Autism has symptoms can hardly be recognized in the early stages of the disease, and it affects the child's
mental health on the long term. Autism can be identified by parents monitoring to the child and diagnosed
by psychiatrists using an international standard checklist. The checklist questions should be answered by
the parent and psychiatrist to determine the risk level of autism (high, medium, or low risk). It is hard for
parents to monitor more than 20 child's behaviours at the same time regardless lack of accuracy for
answering on most of these questions. We propose a system for monitoring autistic child behaviours by
analysing accelerometer data collected from wearable mobile device. The behaviours are recognized by
using a novel algorithm called DTWDir that based on calculating displacement and direction between two
signals. DTWDir is evaluated by comparing it to KNN, classical Dynamic Time Warping (DTW), and One
Dollar Recognition ($1) algorithms. The results show that DTWDir accuracy is higher than the others.
Screening Children with Autism by Developing Smart Toy CarsIJRES Journal
Autistic spectrum disorders are categorized under developmental disorders and lead into a wide range of symptoms in the patients. They are referred to as autism spectrum, since the intensity and the extent of their symptoms vary person to person based on disorder's intensity. In this study, a smart toy car was developed for screening children with autism. In making the toy car, field-programmable gate array was exploited which is an integrated circuit capable of high-speed planning. Appropriate selected features were the ones that identified repetitive and stereotypical motions in movement, which are normally comprised of 9 features. Finally, in the most appropriate conditions, the classification of aforementioned features was carried out using support vector machine algorithm with polynomial kernel with an order of 5. Autistic children screening was carried out with an accuracy of 100%
This study investigated whether 3- to 4.5-month-old infants build causal models linking their own movements to external effects. Infants were exposed to a phase where moving one arm triggered an audiovisual animation, followed by a phase where this contingency was discontinued. The study measured infants' movement patterns and neural responses (EEG). It was hypothesized that if infants formed a causal model, they would show a violation of expectation response (mismatch negativity in EEG) upon the contingency's discontinuation, as well as an initial increase and subsequent decrease in movement (an extinction burst) as their model updated. Results showed infants who demonstrated the EEG response also showed a more pronounced extinction burst, indicating the emergence of a causal sense of agency in
Data Analytics Project proposal: Smart home based ambient assisted living - D...Tarun Swarup
In Ambient Assisted Living environments, monitoring the elderly population can detect a wide range of environmental and user-specific parameters such as daily activities, a regular period of inactivity, usual behavioural patterns and other basic routines. The prime goal of this proposal is to experiment the anomaly detection methods and clustering techniques such as K-means, local outlier factor, K-nearest, DBSCAN and CURE on data and determine the most efficient and accurate method among all.
The Differences in Coordination between Children with ADHD MikeEly930
The Differences in Coordination between Children
with ADHD and Healthy Children Based on Two-
way ANOVA Analysis
jiahai Liu1 gao Yang1 fangzhong Xu minyan Zhou
1:Zhejiang University City College, Hangzhou, China
1
2:Zhejiang Palit Hospital, Hangzhou,China
Abstract—This study attempts to invest the differences in
movement coordination between children with ADHD and
healthy children using two-way ANOVA. The experimental tasks
are divided into simple task and complex task. The goal of the
experiments is to study the interaction in hand movements’
rhythm, accuracy and error key response between task difficulty
and subjects grouping while doing visual-motor integration tasks.
The results show that: (1) There are no significant differences in
all parameters except error number within group differences. (2)
There are significant differences in all parameters except correct
response time between group differences.
Keywords-task difficulty; ADHD; grouping; movement
coordination
I. INTRODUCTION
Attention deficit hyperactivity disorder (ADHD) is very
common in clinical, it is one of the most widely studied
diseases in child psychiatry[1-5]. In recent years, it ranks in the
first or second place in the child-patient cases, causing the
whole society's attention on children with ADHD.
Children with ADHD often accompany with developmental
disabilities, the most common is developmental coordination
disorder (DCD). This is a special developmental disorder, is
characterized by obvious damage in the motion coordination[4-
5]. Although motor coordination disorder and ADHD are two
different developmental disorders, but the studies find that
ADHD often accompanies with motor coordination
disorders[6,7]. Although many of the motor need the
participation of many senses, but visual sense is the most
important. Visual - motor integration is developed firstly in
sensorimotor integration, coordination between visual
perception and hand movements reflects the coordination and
unification between visual perception and activities[8].
At present, continuous performance tests (CPT) are used to
evaluate attention disorder among children. A series of
numbers or characters are showed on a computer monitor as
stimulants in a CPT, subjects are required to make response to
certain target, the reaction results are used to evaluate attention
deficit. Error number in the tests is used to reflect the subjects’
attention deficit, false number reflects the impulse of subjects.
As that, the objectivity of evaluation has been improved greatly
and the accuracy of clinical diagnostic is higher. Andrew L.
Cohen [4] considers that CPT has moderate reliability and
validity in the diagnosis of ADHD. Wang shu-yu[5], finds that
in auditory continuity tests and audio-visual continuity test, the
reaction time of children with all ADHD subtypes are longer
than those of control group, and the number of misstatements
and omissions are significantly ...
Predictive learning of sensorimotor information is hypothesized to be the underlying mechanism that drives cognitive development. As infants learn to minimize the prediction error between their sensory feedback and predictions, they develop various cognitive abilities sequentially through two processes:
1) Learning the relationship between their actions and sensory consequences to develop self-other cognition and goal-directed action.
2) Producing imitative actions in response to others' actions, which allows for the development of imitation, altruistic behavior, and social cognition.
This hypothesis is supported by computational models that show how predictive learning can account for the emergence of skills like self-other discrimination, mirror neuron systems, imitation, and hierarchical representations of goal-directed action in infants
A computational intelligent analysis of autism spectrum disorder using machin...IAESIJAI
Children between the ages of 12 and 24 months who have autism spectrum disorder (ASD) experience abnormalities in the brain that result in undesirable symptoms. Children with ASD struggle to comprehend what others are trying to say and or feel, and they experience extreme anxiety in social situations. Additionally, they have a hard time making friends and even living independently. The defective genes, which control the brain and govern how brain cells communicate with one another, are the primary cause of ASD because they alter brain function. Our primary goal is to assist therapists and parents of children with ASD in using current technologies, such as human intelligence and artificial intelligence, to treat ASD and assist those youngsters in obtaining better social interaction and societal integration. For the purpose of doing an early analysis of ASD, the data is divided into the following three categories: age, gender, and jaundice symptoms. The performance of machine learning algorithms can be influenced by a variety of factors, such as the size of the dataset and quality of the dataset, the choice of features, and the tuning of hyper-parameters. In this work, the support vector machine (SVM) yields 96% as the highest classification accuracy.
This study had three objectives: 1) to determine if an activity-tracking device could increase physical activity, 2) to test if the Theory of Planned Behavior could predict physical activity, and 3) to examine the relationship between physical activity and mobile/transport time. The study found that physical activity significantly increased when wearing the tracking device but the TPB did not predict activity levels. Physical activity also showed no relationship with mobile/transport time. The study supports using tracking devices to encourage physical activity but found limitations with applying the TPB in a technology context and sustaining long-term behavior change.
Screening Children with Autism by Developing Smart Toy CarsIJRES Journal
Autistic spectrum disorders are categorized under developmental disorders and lead into a wide range of symptoms in the patients. They are referred to as autism spectrum, since the intensity and the extent of their symptoms vary person to person based on disorder's intensity. In this study, a smart toy car was developed for screening children with autism. In making the toy car, field-programmable gate array was exploited which is an integrated circuit capable of high-speed planning. Appropriate selected features were the ones that identified repetitive and stereotypical motions in movement, which are normally comprised of 9 features. Finally, in the most appropriate conditions, the classification of aforementioned features was carried out using support vector machine algorithm with polynomial kernel with an order of 5. Autistic children screening was carried out with an accuracy of 100%
This study investigated whether 3- to 4.5-month-old infants build causal models linking their own movements to external effects. Infants were exposed to a phase where moving one arm triggered an audiovisual animation, followed by a phase where this contingency was discontinued. The study measured infants' movement patterns and neural responses (EEG). It was hypothesized that if infants formed a causal model, they would show a violation of expectation response (mismatch negativity in EEG) upon the contingency's discontinuation, as well as an initial increase and subsequent decrease in movement (an extinction burst) as their model updated. Results showed infants who demonstrated the EEG response also showed a more pronounced extinction burst, indicating the emergence of a causal sense of agency in
Data Analytics Project proposal: Smart home based ambient assisted living - D...Tarun Swarup
In Ambient Assisted Living environments, monitoring the elderly population can detect a wide range of environmental and user-specific parameters such as daily activities, a regular period of inactivity, usual behavioural patterns and other basic routines. The prime goal of this proposal is to experiment the anomaly detection methods and clustering techniques such as K-means, local outlier factor, K-nearest, DBSCAN and CURE on data and determine the most efficient and accurate method among all.
The Differences in Coordination between Children with ADHD MikeEly930
The Differences in Coordination between Children
with ADHD and Healthy Children Based on Two-
way ANOVA Analysis
jiahai Liu1 gao Yang1 fangzhong Xu minyan Zhou
1:Zhejiang University City College, Hangzhou, China
1
2:Zhejiang Palit Hospital, Hangzhou,China
Abstract—This study attempts to invest the differences in
movement coordination between children with ADHD and
healthy children using two-way ANOVA. The experimental tasks
are divided into simple task and complex task. The goal of the
experiments is to study the interaction in hand movements’
rhythm, accuracy and error key response between task difficulty
and subjects grouping while doing visual-motor integration tasks.
The results show that: (1) There are no significant differences in
all parameters except error number within group differences. (2)
There are significant differences in all parameters except correct
response time between group differences.
Keywords-task difficulty; ADHD; grouping; movement
coordination
I. INTRODUCTION
Attention deficit hyperactivity disorder (ADHD) is very
common in clinical, it is one of the most widely studied
diseases in child psychiatry[1-5]. In recent years, it ranks in the
first or second place in the child-patient cases, causing the
whole society's attention on children with ADHD.
Children with ADHD often accompany with developmental
disabilities, the most common is developmental coordination
disorder (DCD). This is a special developmental disorder, is
characterized by obvious damage in the motion coordination[4-
5]. Although motor coordination disorder and ADHD are two
different developmental disorders, but the studies find that
ADHD often accompanies with motor coordination
disorders[6,7]. Although many of the motor need the
participation of many senses, but visual sense is the most
important. Visual - motor integration is developed firstly in
sensorimotor integration, coordination between visual
perception and hand movements reflects the coordination and
unification between visual perception and activities[8].
At present, continuous performance tests (CPT) are used to
evaluate attention disorder among children. A series of
numbers or characters are showed on a computer monitor as
stimulants in a CPT, subjects are required to make response to
certain target, the reaction results are used to evaluate attention
deficit. Error number in the tests is used to reflect the subjects’
attention deficit, false number reflects the impulse of subjects.
As that, the objectivity of evaluation has been improved greatly
and the accuracy of clinical diagnostic is higher. Andrew L.
Cohen [4] considers that CPT has moderate reliability and
validity in the diagnosis of ADHD. Wang shu-yu[5], finds that
in auditory continuity tests and audio-visual continuity test, the
reaction time of children with all ADHD subtypes are longer
than those of control group, and the number of misstatements
and omissions are significantly ...
Predictive learning of sensorimotor information is hypothesized to be the underlying mechanism that drives cognitive development. As infants learn to minimize the prediction error between their sensory feedback and predictions, they develop various cognitive abilities sequentially through two processes:
1) Learning the relationship between their actions and sensory consequences to develop self-other cognition and goal-directed action.
2) Producing imitative actions in response to others' actions, which allows for the development of imitation, altruistic behavior, and social cognition.
This hypothesis is supported by computational models that show how predictive learning can account for the emergence of skills like self-other discrimination, mirror neuron systems, imitation, and hierarchical representations of goal-directed action in infants
A computational intelligent analysis of autism spectrum disorder using machin...IAESIJAI
Children between the ages of 12 and 24 months who have autism spectrum disorder (ASD) experience abnormalities in the brain that result in undesirable symptoms. Children with ASD struggle to comprehend what others are trying to say and or feel, and they experience extreme anxiety in social situations. Additionally, they have a hard time making friends and even living independently. The defective genes, which control the brain and govern how brain cells communicate with one another, are the primary cause of ASD because they alter brain function. Our primary goal is to assist therapists and parents of children with ASD in using current technologies, such as human intelligence and artificial intelligence, to treat ASD and assist those youngsters in obtaining better social interaction and societal integration. For the purpose of doing an early analysis of ASD, the data is divided into the following three categories: age, gender, and jaundice symptoms. The performance of machine learning algorithms can be influenced by a variety of factors, such as the size of the dataset and quality of the dataset, the choice of features, and the tuning of hyper-parameters. In this work, the support vector machine (SVM) yields 96% as the highest classification accuracy.
This study had three objectives: 1) to determine if an activity-tracking device could increase physical activity, 2) to test if the Theory of Planned Behavior could predict physical activity, and 3) to examine the relationship between physical activity and mobile/transport time. The study found that physical activity significantly increased when wearing the tracking device but the TPB did not predict activity levels. Physical activity also showed no relationship with mobile/transport time. The study supports using tracking devices to encourage physical activity but found limitations with applying the TPB in a technology context and sustaining long-term behavior change.
Analysis & Detection of autism spectrum disorder using ML (3)_RKS.pptxssuser24292c
This document presents a machine learning model for detecting autism spectrum disorder (ASD) in early developmental stages. It discusses ASD symptoms and impacts, with the goal of using ML to help streamline diagnosis. Five ML classifiers were trained on a dataset containing features like age, sex and symptoms. Logistic regression achieved the highest accuracy. The project aims to provide an automated early detection tool, as current diagnosis is time-consuming. However, more data is needed to build an accurate model. In summary, the document describes a ML model for early ASD detection, compares classifier performances, and discusses the need for more data to improve the model.
Development of Cognitive Instruments in Epidemiology Using Asyncronous MethodsAJHSSR Journal
ABSTRACT :The purpose of this research is to find out the development of cognitive instruments for
community empowerment in the field of epidemiology using the asynchronous method. This research method
uses a quantitative descriptive method. The results of the study found that learning using asynchronous elearning was inappropriate for students with low cognitive levels. So that before learning, a pre-test was carried
out, from the results of filling in the items it could be seen that the cognitive level of students was high, low or
medium. For high cognitive levels, it can still be given. But for moderate or low cognitive levels, it is necessary
to do a good learning design, so that these students can follow well the entire learning process. Conclusion of
this study From the results of several research results, it can be concluded that learning using asynchronous e
learning is less effective, especially students' low cognitive levels. Requires psychological support from the
environment. because there is no direct interaction, so that between students do not know each other. especially
the support from the tutor/teacher. Students learn according to their respective work settings, so that they learn
independently. It is necessary to do a mature design before the learning begins. So that learning can have a good
impact, namely being able to improve the performance of the learner.
KEYWORDS: Cognitive development, instruments, asynchronous
Schizophrenia is a mental illness with a very bad impact on sufferers, attacking the part of human brain that disables the ability to think clearly. In 2018, Rustam and Rampisela classified Schizophrenia by using Northwestern University Schizophrenia Data, based on 66 variables consisting of group, demographic, and questionnaires statistics, based on the scale for the assessment of negative symptoms (SANS), and scale for the assessment of positive symptoms (SAS), and then classifiers that used are SVM with Gaussian kernel and Twin SVM with linear and Gaussian kernel. Furthermore, this research is novel based on the use of random forest as a classifier, in order to predict Schizophrenia. The result obtained is reported in percentage of accuracy, both in training and testing of random forest, which was 100%. This classification, therefore, shows the best value in contrast with prior methods, even though only 40% of training data set was used. This is very important, especially in the cases of rare disease, including schizophrenia.
Depression prognosis using natural language processing and machine learning ...IJECEIAES
Depression is an acute problem throughout the world. Due to worst and prolong depression many people dies in every year. The problem is that most of the people are not concern of the fact that they are suffering from depression. In this research, our aim was to find out whether an individual is depressed or not by analyzing social media status. Therefore, we focused on real data. Our dataset consists of 2,000 sentences, which was collected from different social media platforms Facebook, Twitter, and Instagram. Then, we have performed five data pre-processing approaches for natural language processing (NLP) such as tokenization, removal of stop words, removing empty string, removing punctuations, stemming and lemmatization. For our selected model, we considered that processed data as an input. Finally, we applied six machine learning (ML) classifiers multinomial Naive Bayes (NB), logistic regression, liner support vector classifier, random forest, K-nearest neighbour, and decision tree to achieve better accuracy over our dataset. Among six algorithms, multinomial NB and logistic regression performed well on our dataset and obtained 98% accuracy.
Quantifying the efficacy of ML models at predicting mental health illnessesIRJET Journal
This document summarizes a research study that evaluated the efficacy of machine learning models at predicting mental health illnesses like depression. The study collected self-reported survey data from participants over two weeks to quantify emotional variability and depressive symptoms. It then used this data to train and compare the accuracy of logistic regression, random forest, and multi-layer perceptron models against a baseline model and participants' Beck Depression Inventory scores. The preliminary results found a positive correlation between emotional variability, amount of labeled data/features, and model accuracy. The random forest model was most accurate at predicting depression incidence compared to the other models and baseline. The research aims to assess the benefits and limitations of using ML to detect mental health issues like depression.
Autism Spectrum Disorders Gait Identification Using Ground Reaction ForcesTELKOMNIKA JOURNAL
Autism spectrum disorders (ASD) are a permanent neurodevelopmental disorder that can be
identified during the first few years of life and are currently associated with the abnormal walking pattern.
Earlier identification of this pervasive disorder could provide assistance in diagnosis and establish rapid
quantitative clinical judgment. This paper presents an automated approach which can be applied to identify
ASD gait patterns using three-dimensional (3D) ground reaction forces (GRF). The study involved
classification of gait patterns of children with ASD and typical healthy children. The GRF data were
obtained using two force plates during self-determined barefoot walking. Time-series parameterization
techniques were applied to the GRF waveforms to extract the important gait features. The most dominant
and correct features for characterizing ASD gait were selected using statistical between-group tests and
stepwise discriminant analysis (SWDA). The selected features were grouped into two groups which served
as two input datasets to the k-nearest neighbor (KNN) classifier. This study demonstrates that the 3D GRF
gait features selected using SWDA are reliable to be used in the identification of ASD gait using KNN
classifier with 83.33% performance accuracy.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASESpijans
Pediatric emergency cases need rapid systems that measure vital body parameters data, analyze and
categorize emergency cases for precise action. Current systems use manual examination resulting in
delayed medication, death, or other severe medical conditions.In this paper, we propose a Internet of
Things (IoT) based model, created using Balena fin with Raspberry pi compute module. It is used for
determining emergency cases, in pediatric section, specifically the triage section. It is later tested using
hospital data that represents the vital parameters in pediatric. Our approach entails designing and setting
up the hardware and software infrastructure, to accommodate data via Bluetooth protocol, and transmit it
to the cloud server database via Message Queuing Telemetry Transport (MQTT). Later, we perform
machine learning on the data by training a model and finally develop a Plotly Dash analytical application
integrating the model for visualization near real-time.Findings show that emergency cases are detected
using vital body parameters which include the body temperature, oxygen levels, heart rate and the age. The
model indicates a 97% accuracy.In conclusion, children’s emergency cases are detected in time using IoT
gadgets and machine learning classification.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASESpijans
This document proposes an IoT-based model for identifying pediatric emergency cases using vital body parameters data collected from Bluetooth sensors. The model uses a Raspberry Pi device to collect temperature, oxygen, heart rate, and blood pressure data and transmit it via MQTT to an AWS cloud database. A machine learning model is trained on historical hospital data and achieves 97% accuracy in classifying cases as emergency, non-emergency, or moderate emergency. The model provides a way to rapidly identify pediatric emergency cases using wireless sensors and cloud-based analytics.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASESamsjournal
Pediatric emergency cases need rapid systems that measure vital body parameters data, analyze and
categorize emergency cases for precise action. Current systems use manual examination resulting in
delayed medication, death, or other severe medical conditions.In this paper, we propose a Internet of
Things (IoT) based model, created using Balena fin with Raspberry pi compute module. It is used for
determining emergency cases, in pediatric section, specifically the triage section. It is later tested using
hospital data that represents the vital parameters in pediatric. Our approach entails designing and setting
up the hardware and software infrastructure, to accommodate data via Bluetooth protocol, and transmit it
to the cloud server database via Message Queuing Telemetry Transport (MQTT). Later, we perform
machine learning on the data by training a model and finally develop a Plotly Dash analytical application
integrating the model for visualization near real-time.Findings show that emergency cases are detected
using vital body parameters which include the body temperature, oxygen levels, heart rate and the age. The
model indicates a 97% accuracy.In conclusion, children’s emergency cases are detected in time using IoT
gadgets and machine learning classification.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASES pijans
Pediatric emergency cases need rapid systems that measure vital body parameters data, analyze and
categorize emergency cases for precise action. Current systems use manual examination resulting in
delayed medication, death, or other severe medical conditions.In this paper, we propose a Internet of
Things (IoT) based model, created using Balena fin with Raspberry pi compute module. It is used for
determining emergency cases, in pediatric section, specifically the triage section. It is later tested using
hospital data that represents the vital parameters in pediatric. Our approach entails designing and setting
up the hardware and software infrastructure, to accommodate data via Bluetooth protocol, and transmit it
to the cloud server database via Message Queuing Telemetry Transport (MQTT)
IRJET-Artificial Neural Network and Fuzzy Logic Approach to Diagnose Autism S...IRJET Journal
This document describes a study that uses an artificial neural network and fuzzy logic approach to diagnose autism spectrum disorder early. The study aims to address the complexity of autism symptoms and the accuracy of information collected from parents, which current diagnosis methods struggle with. It proposes using a combination of artificial neural network and fuzzy logic to establish a system for early autism diagnosis. This approach has been shown to be effective in other fields like finance, geology and medicine. The study generates a sample dataset to test the proposed diagnosis model, which uses fuzzy inputs and weights to account for uncertainty in the data. The results indicate this approach has potential for use in autism research and diagnosis support systems.
This proposal aims to use deep learning algorithms and data from wearable devices, multi-omic data, and a cognitive city to predict adverse health outcomes in healthy populations. Specifically, it will use LSTM and GAN models on sparse and dense time-series data to predict risks of conditions like diabetes, hypertension, heart attacks, and adverse pregnancy outcomes. The proposal will apply cutting-edge methods like deep learning to take advantage of improvements in feature extraction and accuracy over traditional machine learning.
This proposal aims to use deep learning algorithms and data from wearable devices, multi-omic data, and a cognitive city to predict adverse health outcomes in healthy populations. Specifically, it will use LSTM and GAN models on sparse and dense time-series data to predict risks of conditions like diabetes, hypertension, heart attacks, and adverse pregnancy outcomes. The proposal will apply cutting-edge methods like deep learning to take advantage of improvements in feature extraction and accuracy over traditional machine learning.
Bayesain Hypothesis of Selective Attention - Raw 2011 posterGiacomo Veneri
The aim of the study is to understand the process of target averaging during the selection process. We analyzed the probability to select the target after a fixation outside ROIs from the duration of fixations and the distance to the target. We aimed to respond to the question “is it possible to predict the selected area?” . In this study we tested the presence of information in non-ROI fixation data about the occurrence of a target at the next saccade. A classification algorithm was trained to predict the target vs. non-target outcome (dependent variable) of a saccade from summary statistics of fixation data (covariates). We claim that significantly accurate predictions are substantial evidence to support the hypothesis of "presence of information".
8 2008-normative data for the letter cancellation task in school childrenElsa von Licy
The document presents normative data for the letter-cancellation task, a psychomotor performance test, in 819 Indian school children aged 9-16 years. Both age and sex influenced performance, with scores increasing with age and higher in females. Regression models were used to develop normative data tables stratified by age and sex, allowing for quantitative evaluation and wider clinical use of the letter-cancellation task in India.
Recent advances and challenges of digital mental healthcareYoon Sup Choi
This document discusses research analyzing the relationship between mobile phone location sensor data and measures of depressive symptom severity. The research replicated a previous study finding significant correlations between several GPS-derived features (location variance, entropy, circadian movement) and scores on the PHQ-9 depression scale. These relationships were stronger when analyzing weekend versus weekday GPS data. GPS features predicted PHQ-9 scores up to 10 weeks later, suggesting they may serve as early warning signals of depression. The findings provide further evidence that passively collected GPS data from smartphones can reliably predict depressive symptom severity.
TOWARDS CEREBRAL PALSY DIAGNOSIS: ANONTOLOGY BASED APPROACHijseajournal
Cerebral Palsy (CP) is one of the most complicated disabilities which is a permanent motor disorder
causing mental and physical disabilities. Different reports published by different health organizations
asked for researches on CP disability in order to improve diagnosis. Globally, there are different
researches conducted to improve CP diagnosis, but most of those studies do not diagnose CP in children’s
early ages, which limit the treatment impact. This paper report on a research conducted to develop an
ontology-based approach to diagnose children with CP in early ages. In this paper, Ontology was used to
represent CP domain. Then, a set of manually built rules have been optimized through a knowledge-based
survey to be used in CP diagnosis. The proposed approach improves CP diagnosis and by consequence
positively reflects physical therapy treatment. The proposed approach was evaluated by a real dataset
consisted of 70 pre-diagnosed cases, where 84% correctly diagnosed.
This document summarizes a study on psychiatric symptoms among children with congenital heart disease. The study aimed to examine depressive and anxiety symptoms as well as neurocognitive deficits in children with congenital heart disease compared to controls. It found that children with congenital heart disease performed significantly worse on tests of cognitive functioning and had higher levels of depressive and anxiety symptoms than controls. Common psychiatric diagnoses among the children with heart disease included adjustment disorder and depression. The results suggest children with congenital heart disease are at increased risk for psychological and cognitive issues.
Users Approach on Providing Feedback for Smart Home Devices – Phase IIijujournal
Smart Home technology has accomplished extraordinary success in making individuals' lives more straightforward and relaxing. Technology has recently brought about numerous savvy and refined frame works that advanced clever living innovation. In this paper, we will investigate the behavioral intention of user's approach to providing feedback for smart home devices. We will conduct an online survey for a sample of three to five students selected by simple random sampling to study the user's motto for giving feedback on smart home devices and their expectations. We have observed that most users are ready to actively share their input on smart home devices to improve the product's service and quality to fulfill the user’s needs and make their lives easier.
Users Approach on Providing Feedback for Smart Home Devices – Phase IIijujournal
Smart Home technology has accomplished extraordinary success in making individuals' lives more
straightforward and relaxing. Technology has recently brought about numerous savvy and refined frame
works that advanced clever living innovation. In this paper, we will investigate the behavioral intention of
user's approach to providing feedback for smart home devices. We will conduct an online survey for a
sample of three to five students selected by simple random sampling to study the user's motto for giving
feedback on smart home devices and their expectations. We have observed that most users are ready to
actively share their input on smart home devices to improve the product's service and quality to fulfill the
user’s needs and make their lives easier.
More Related Content
Similar to DTWDIR: AN ENHANCED DTW ALGORITHM FOR AUTISTIC CHILD BEHAVIOUR MONITORING
Analysis & Detection of autism spectrum disorder using ML (3)_RKS.pptxssuser24292c
This document presents a machine learning model for detecting autism spectrum disorder (ASD) in early developmental stages. It discusses ASD symptoms and impacts, with the goal of using ML to help streamline diagnosis. Five ML classifiers were trained on a dataset containing features like age, sex and symptoms. Logistic regression achieved the highest accuracy. The project aims to provide an automated early detection tool, as current diagnosis is time-consuming. However, more data is needed to build an accurate model. In summary, the document describes a ML model for early ASD detection, compares classifier performances, and discusses the need for more data to improve the model.
Development of Cognitive Instruments in Epidemiology Using Asyncronous MethodsAJHSSR Journal
ABSTRACT :The purpose of this research is to find out the development of cognitive instruments for
community empowerment in the field of epidemiology using the asynchronous method. This research method
uses a quantitative descriptive method. The results of the study found that learning using asynchronous elearning was inappropriate for students with low cognitive levels. So that before learning, a pre-test was carried
out, from the results of filling in the items it could be seen that the cognitive level of students was high, low or
medium. For high cognitive levels, it can still be given. But for moderate or low cognitive levels, it is necessary
to do a good learning design, so that these students can follow well the entire learning process. Conclusion of
this study From the results of several research results, it can be concluded that learning using asynchronous e
learning is less effective, especially students' low cognitive levels. Requires psychological support from the
environment. because there is no direct interaction, so that between students do not know each other. especially
the support from the tutor/teacher. Students learn according to their respective work settings, so that they learn
independently. It is necessary to do a mature design before the learning begins. So that learning can have a good
impact, namely being able to improve the performance of the learner.
KEYWORDS: Cognitive development, instruments, asynchronous
Schizophrenia is a mental illness with a very bad impact on sufferers, attacking the part of human brain that disables the ability to think clearly. In 2018, Rustam and Rampisela classified Schizophrenia by using Northwestern University Schizophrenia Data, based on 66 variables consisting of group, demographic, and questionnaires statistics, based on the scale for the assessment of negative symptoms (SANS), and scale for the assessment of positive symptoms (SAS), and then classifiers that used are SVM with Gaussian kernel and Twin SVM with linear and Gaussian kernel. Furthermore, this research is novel based on the use of random forest as a classifier, in order to predict Schizophrenia. The result obtained is reported in percentage of accuracy, both in training and testing of random forest, which was 100%. This classification, therefore, shows the best value in contrast with prior methods, even though only 40% of training data set was used. This is very important, especially in the cases of rare disease, including schizophrenia.
Depression prognosis using natural language processing and machine learning ...IJECEIAES
Depression is an acute problem throughout the world. Due to worst and prolong depression many people dies in every year. The problem is that most of the people are not concern of the fact that they are suffering from depression. In this research, our aim was to find out whether an individual is depressed or not by analyzing social media status. Therefore, we focused on real data. Our dataset consists of 2,000 sentences, which was collected from different social media platforms Facebook, Twitter, and Instagram. Then, we have performed five data pre-processing approaches for natural language processing (NLP) such as tokenization, removal of stop words, removing empty string, removing punctuations, stemming and lemmatization. For our selected model, we considered that processed data as an input. Finally, we applied six machine learning (ML) classifiers multinomial Naive Bayes (NB), logistic regression, liner support vector classifier, random forest, K-nearest neighbour, and decision tree to achieve better accuracy over our dataset. Among six algorithms, multinomial NB and logistic regression performed well on our dataset and obtained 98% accuracy.
Quantifying the efficacy of ML models at predicting mental health illnessesIRJET Journal
This document summarizes a research study that evaluated the efficacy of machine learning models at predicting mental health illnesses like depression. The study collected self-reported survey data from participants over two weeks to quantify emotional variability and depressive symptoms. It then used this data to train and compare the accuracy of logistic regression, random forest, and multi-layer perceptron models against a baseline model and participants' Beck Depression Inventory scores. The preliminary results found a positive correlation between emotional variability, amount of labeled data/features, and model accuracy. The random forest model was most accurate at predicting depression incidence compared to the other models and baseline. The research aims to assess the benefits and limitations of using ML to detect mental health issues like depression.
Autism Spectrum Disorders Gait Identification Using Ground Reaction ForcesTELKOMNIKA JOURNAL
Autism spectrum disorders (ASD) are a permanent neurodevelopmental disorder that can be
identified during the first few years of life and are currently associated with the abnormal walking pattern.
Earlier identification of this pervasive disorder could provide assistance in diagnosis and establish rapid
quantitative clinical judgment. This paper presents an automated approach which can be applied to identify
ASD gait patterns using three-dimensional (3D) ground reaction forces (GRF). The study involved
classification of gait patterns of children with ASD and typical healthy children. The GRF data were
obtained using two force plates during self-determined barefoot walking. Time-series parameterization
techniques were applied to the GRF waveforms to extract the important gait features. The most dominant
and correct features for characterizing ASD gait were selected using statistical between-group tests and
stepwise discriminant analysis (SWDA). The selected features were grouped into two groups which served
as two input datasets to the k-nearest neighbor (KNN) classifier. This study demonstrates that the 3D GRF
gait features selected using SWDA are reliable to be used in the identification of ASD gait using KNN
classifier with 83.33% performance accuracy.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASESpijans
Pediatric emergency cases need rapid systems that measure vital body parameters data, analyze and
categorize emergency cases for precise action. Current systems use manual examination resulting in
delayed medication, death, or other severe medical conditions.In this paper, we propose a Internet of
Things (IoT) based model, created using Balena fin with Raspberry pi compute module. It is used for
determining emergency cases, in pediatric section, specifically the triage section. It is later tested using
hospital data that represents the vital parameters in pediatric. Our approach entails designing and setting
up the hardware and software infrastructure, to accommodate data via Bluetooth protocol, and transmit it
to the cloud server database via Message Queuing Telemetry Transport (MQTT). Later, we perform
machine learning on the data by training a model and finally develop a Plotly Dash analytical application
integrating the model for visualization near real-time.Findings show that emergency cases are detected
using vital body parameters which include the body temperature, oxygen levels, heart rate and the age. The
model indicates a 97% accuracy.In conclusion, children’s emergency cases are detected in time using IoT
gadgets and machine learning classification.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASESpijans
This document proposes an IoT-based model for identifying pediatric emergency cases using vital body parameters data collected from Bluetooth sensors. The model uses a Raspberry Pi device to collect temperature, oxygen, heart rate, and blood pressure data and transmit it via MQTT to an AWS cloud database. A machine learning model is trained on historical hospital data and achieves 97% accuracy in classifying cases as emergency, non-emergency, or moderate emergency. The model provides a way to rapidly identify pediatric emergency cases using wireless sensors and cloud-based analytics.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASESamsjournal
Pediatric emergency cases need rapid systems that measure vital body parameters data, analyze and
categorize emergency cases for precise action. Current systems use manual examination resulting in
delayed medication, death, or other severe medical conditions.In this paper, we propose a Internet of
Things (IoT) based model, created using Balena fin with Raspberry pi compute module. It is used for
determining emergency cases, in pediatric section, specifically the triage section. It is later tested using
hospital data that represents the vital parameters in pediatric. Our approach entails designing and setting
up the hardware and software infrastructure, to accommodate data via Bluetooth protocol, and transmit it
to the cloud server database via Message Queuing Telemetry Transport (MQTT). Later, we perform
machine learning on the data by training a model and finally develop a Plotly Dash analytical application
integrating the model for visualization near real-time.Findings show that emergency cases are detected
using vital body parameters which include the body temperature, oxygen levels, heart rate and the age. The
model indicates a 97% accuracy.In conclusion, children’s emergency cases are detected in time using IoT
gadgets and machine learning classification.
INTERNET OF THINGS BASED MODEL FOR IDENTIFYING PEDIATRIC EMERGENCY CASES pijans
Pediatric emergency cases need rapid systems that measure vital body parameters data, analyze and
categorize emergency cases for precise action. Current systems use manual examination resulting in
delayed medication, death, or other severe medical conditions.In this paper, we propose a Internet of
Things (IoT) based model, created using Balena fin with Raspberry pi compute module. It is used for
determining emergency cases, in pediatric section, specifically the triage section. It is later tested using
hospital data that represents the vital parameters in pediatric. Our approach entails designing and setting
up the hardware and software infrastructure, to accommodate data via Bluetooth protocol, and transmit it
to the cloud server database via Message Queuing Telemetry Transport (MQTT)
IRJET-Artificial Neural Network and Fuzzy Logic Approach to Diagnose Autism S...IRJET Journal
This document describes a study that uses an artificial neural network and fuzzy logic approach to diagnose autism spectrum disorder early. The study aims to address the complexity of autism symptoms and the accuracy of information collected from parents, which current diagnosis methods struggle with. It proposes using a combination of artificial neural network and fuzzy logic to establish a system for early autism diagnosis. This approach has been shown to be effective in other fields like finance, geology and medicine. The study generates a sample dataset to test the proposed diagnosis model, which uses fuzzy inputs and weights to account for uncertainty in the data. The results indicate this approach has potential for use in autism research and diagnosis support systems.
This proposal aims to use deep learning algorithms and data from wearable devices, multi-omic data, and a cognitive city to predict adverse health outcomes in healthy populations. Specifically, it will use LSTM and GAN models on sparse and dense time-series data to predict risks of conditions like diabetes, hypertension, heart attacks, and adverse pregnancy outcomes. The proposal will apply cutting-edge methods like deep learning to take advantage of improvements in feature extraction and accuracy over traditional machine learning.
This proposal aims to use deep learning algorithms and data from wearable devices, multi-omic data, and a cognitive city to predict adverse health outcomes in healthy populations. Specifically, it will use LSTM and GAN models on sparse and dense time-series data to predict risks of conditions like diabetes, hypertension, heart attacks, and adverse pregnancy outcomes. The proposal will apply cutting-edge methods like deep learning to take advantage of improvements in feature extraction and accuracy over traditional machine learning.
Bayesain Hypothesis of Selective Attention - Raw 2011 posterGiacomo Veneri
The aim of the study is to understand the process of target averaging during the selection process. We analyzed the probability to select the target after a fixation outside ROIs from the duration of fixations and the distance to the target. We aimed to respond to the question “is it possible to predict the selected area?” . In this study we tested the presence of information in non-ROI fixation data about the occurrence of a target at the next saccade. A classification algorithm was trained to predict the target vs. non-target outcome (dependent variable) of a saccade from summary statistics of fixation data (covariates). We claim that significantly accurate predictions are substantial evidence to support the hypothesis of "presence of information".
8 2008-normative data for the letter cancellation task in school childrenElsa von Licy
The document presents normative data for the letter-cancellation task, a psychomotor performance test, in 819 Indian school children aged 9-16 years. Both age and sex influenced performance, with scores increasing with age and higher in females. Regression models were used to develop normative data tables stratified by age and sex, allowing for quantitative evaluation and wider clinical use of the letter-cancellation task in India.
Recent advances and challenges of digital mental healthcareYoon Sup Choi
This document discusses research analyzing the relationship between mobile phone location sensor data and measures of depressive symptom severity. The research replicated a previous study finding significant correlations between several GPS-derived features (location variance, entropy, circadian movement) and scores on the PHQ-9 depression scale. These relationships were stronger when analyzing weekend versus weekday GPS data. GPS features predicted PHQ-9 scores up to 10 weeks later, suggesting they may serve as early warning signals of depression. The findings provide further evidence that passively collected GPS data from smartphones can reliably predict depressive symptom severity.
TOWARDS CEREBRAL PALSY DIAGNOSIS: ANONTOLOGY BASED APPROACHijseajournal
Cerebral Palsy (CP) is one of the most complicated disabilities which is a permanent motor disorder
causing mental and physical disabilities. Different reports published by different health organizations
asked for researches on CP disability in order to improve diagnosis. Globally, there are different
researches conducted to improve CP diagnosis, but most of those studies do not diagnose CP in children’s
early ages, which limit the treatment impact. This paper report on a research conducted to develop an
ontology-based approach to diagnose children with CP in early ages. In this paper, Ontology was used to
represent CP domain. Then, a set of manually built rules have been optimized through a knowledge-based
survey to be used in CP diagnosis. The proposed approach improves CP diagnosis and by consequence
positively reflects physical therapy treatment. The proposed approach was evaluated by a real dataset
consisted of 70 pre-diagnosed cases, where 84% correctly diagnosed.
This document summarizes a study on psychiatric symptoms among children with congenital heart disease. The study aimed to examine depressive and anxiety symptoms as well as neurocognitive deficits in children with congenital heart disease compared to controls. It found that children with congenital heart disease performed significantly worse on tests of cognitive functioning and had higher levels of depressive and anxiety symptoms than controls. Common psychiatric diagnoses among the children with heart disease included adjustment disorder and depression. The results suggest children with congenital heart disease are at increased risk for psychological and cognitive issues.
Similar to DTWDIR: AN ENHANCED DTW ALGORITHM FOR AUTISTIC CHILD BEHAVIOUR MONITORING (20)
Users Approach on Providing Feedback for Smart Home Devices – Phase IIijujournal
Smart Home technology has accomplished extraordinary success in making individuals' lives more straightforward and relaxing. Technology has recently brought about numerous savvy and refined frame works that advanced clever living innovation. In this paper, we will investigate the behavioral intention of user's approach to providing feedback for smart home devices. We will conduct an online survey for a sample of three to five students selected by simple random sampling to study the user's motto for giving feedback on smart home devices and their expectations. We have observed that most users are ready to actively share their input on smart home devices to improve the product's service and quality to fulfill the user’s needs and make their lives easier.
Users Approach on Providing Feedback for Smart Home Devices – Phase IIijujournal
Smart Home technology has accomplished extraordinary success in making individuals' lives more
straightforward and relaxing. Technology has recently brought about numerous savvy and refined frame
works that advanced clever living innovation. In this paper, we will investigate the behavioral intention of
user's approach to providing feedback for smart home devices. We will conduct an online survey for a
sample of three to five students selected by simple random sampling to study the user's motto for giving
feedback on smart home devices and their expectations. We have observed that most users are ready to
actively share their input on smart home devices to improve the product's service and quality to fulfill the
user’s needs and make their lives easier.
October 2023-Top Cited Articles in IJU.pdfijujournal
International Journal of Ubiquitous Computing (IJU) is a quarterly open access peer-reviewed journal that provides excellent international forum for sharing knowledge and results in theory, methodology and applications of ubiquitous computing. Current information age is witnessing a dramatic use of digital and electronic devices in the workplace and beyond. Ubiquitous Computing presents a rather arduous requirement of robustness, reliability and availability to the end user. Ubiquitous computing has received a significant and sustained research interest in terms of designing and deploying large scale and high performance computational applications in real life. The aim of the journal is to provide a platform to the researchers and practitioners from both academia as well as industry to meet and share cutting-edge development in the field.
ACCELERATION DETECTION OF LARGE (PROBABLY) PRIME NUMBERSijujournal
This document discusses methods for efficiently generating large prime numbers for use in RSA cryptography. It presents experimental results measuring the time taken to generate prime numbers when trial dividing the starting number by different numbers of initial primes before applying the Miller-Rabin primality test. The optimal number of trial divisions can be estimated as B=E/D, where E is the time for Miller-Rabin test and D is the maximum usefulness of trial division. Experimental results on different sized numbers support dividing by around 20 initial primes as optimal.
A novel integrated approach for handling anomalies in RFID dataijujournal
Radio Frequency Identification (RFID) is a convenient technology employed in various applications. The
success of these RFID applications depends heavily on the quality of the data stream generated by RFID
readers. Due to various anomalies found predominantly in RFID data it limits the widespread adoption of
this technology. Our work is to eliminate the anomalies present in RFID data in an effective manner so that
it can be applied for high end applications. Our approach is a hybrid approach of middleware and
deferred because it is not always possible to remove all anomalies and redundancies in middleware. The
processing of other anomalies is deferred until the query time and cleaned by business rules. Experimental
results show that the proposed approach performs the cleaning in an effective manner compared to the
existing approaches.
UBIQUITOUS HEALTHCARE MONITORING SYSTEM USING INTEGRATED TRIAXIAL ACCELEROMET...ijujournal
Ubiquitous healthcare has become one of the prominent areas of research inorder to address the
challenges encountered in healthcare environment. In contribution to this area, this study developed a
system prototype that recommends diagonostic services based on physiological data collected in real time
from a distant patient. The prototype uses WBAN body sensors to be worn by the individual and an android
smart phone as a personal server. Physiological data is collected and uploaded to a Medical Health
Server (MHS) via GPRS/internet to be analysed. Our implemented prototype monitors the activity, location
and physiological data such as SpO2 and Heart Rate (HR) of the elderly and patients in rehabilitation. The
uploaded information can be accessed in real time by medical practitioners through a web application.
ENHANCING INDEPENDENT SENIOR LIVING THROUGH SMART HOME TECHNOLOGIESijujournal
The population of elderly folks is ballooning worldwide as people live longer. But getting older often
means declining health and trouble living solo. Smart home tech could keep an eye on old folks and get
help quickly when needed so they can stay independent. This paper looks at a system combining wireless
sensors, video watches, automation, resident monitoring, emergency detection, and remote access. Sensors
track health signs, activities, appliance use. Video analytics spot odd stuff like falls. Sensor fusion and
machine learning find normal patterns so wonks can see unhealthy changes and send alerts. Multi-channel
alerts reach caregivers and emergency folks. A LabVIEW can integrate devices and enables local and
remote oversight and can control and handle emergency responses. Benefits seem to be early illness clues,
quick help, less burden on caregivers, and optimized home settings. But will old folks use all this tech? Can
we prove it really helps folks live longer and better? More research on maximizing reliability and
evaluating real-world impacts is needed. But designed thoughtfully, smart homes could may profoundly
improve the aging experience.
HMR LOG ANALYZER: ANALYZE WEB APPLICATION LOGS OVER HADOOP MAPREDUCEijujournal
In today’s Internet world, log file analysis is becoming a necessary task for analyzing the customer’s
behavior in order to improve advertising and sales as well as for datasets like environment, medical,
banking system it is important to analyze the log data to get required knowledge from it. Web mining is the
process of discovering the knowledge from the web data. Log files are getting generated very fast at the
rate of 1-10 Mb/s per machine, a single data center can generate tens of terabytes of log data in a day.
These datasets are huge. In order to analyze such large datasets we need parallel processing system and
reliable data storage mechanism. Virtual database system is an effective solution for integrating the data
but it becomes inefficient for large datasets. The Hadoop framework provides reliable data storage by
Hadoop Distributed File System and MapReduce programming model which is a parallel processing
system for large datasets. Hadoop distributed file system breaks up input data and sends fractions of the
original data to several machines in hadoop cluster to hold blocks of data. This mechanism helps to
process log data in parallel using all the machines in the hadoop cluster and computes result efficiently.
The dominant approach provided by hadoop to “Store first query later”, loads the data to the Hadoop
Distributed File System and then executes queries written in Pig Latin. This approach reduces the response
time as well as the load on to the end system. This paper proposes a log analysis system using Hadoop
MapReduce which will provide accurate results in minimum response time.
SERVICE DISCOVERY – A SURVEY AND COMPARISONijujournal
The document summarizes and compares several major service discovery approaches. It provides an overview of service discovery objectives and techniques, then surveys prominent protocols including SLP, Jini, and UPnP. Each approach is analyzed based on features like service description, discovery architecture, announcement/query mechanisms, and how they handle service usage and dynamic network changes. The comparison aims to identify strengths and limitations to guide future research in improving service discovery.
SIX DEGREES OF SEPARATION TO IMPROVE ROUTING IN OPPORTUNISTIC NETWORKSijujournal
Opportunistic Networks are able to exploit social behavior to create connectivity opportunities. This
paradigm uses pair-wise contacts for routing messages between nodes. In this context we investigated if the
“six degrees of separation” conjecture of small-world networks can be used as a basis to route messages in
Opportunistic Networks. We propose a simple approach for routing that outperforms some popular
protocols in simulations that are carried out with real world traces using ONE simulator. We conclude that
static graph models are not suitable for underlay routing approaches in highly dynamic networks like
Opportunistic Networks without taking account of temporal factors such as time, duration and frequency of
previous encounters.
International Journal of Ubiquitous Computing (IJU)ijujournal
International Journal of Ubiquitous Computing (IJU) is a quarterly open access peer-reviewed journal that provides excellent international forum for sharing knowledge and results in theory, methodology and applications of ubiquitous computing. Current information age is witnessing a dramatic use of digital and electronic devices in the workplace and beyond. Ubiquitous Computing presents a rather arduous requirement of robustness, reliability and availability to the end user. Ubiquitous computing has received a significant and sustained research interest in terms of designing and deploying large scale and high performance computational applications in real life. The aim of the journal is to provide a platform to the researchers and practitioners from both academia as well as industry to meet and share cutting-edge development in the field.
PERVASIVE COMPUTING APPLIED TO THE CARE OF PATIENTS WITH DEMENTIA IN HOMECARE...ijujournal
The aging population and the consequent increase in the incidence of dementias is causing many
challenges to health systems, mainly related to infrastructure, low services quality and high costs. One
solution is to provide the care at house of the patient, through of home care services. However, it is not a
trivial task, since a patient with dementia requires constant care and monitoring from a caregiver, who
suffers physical and emotional overload. In this context, this work presents an modelling for development of
pervasive systems aimed at helping the care of these patients in order to lessen the burden of the caregiver
while the patient continue to receive the necessary care.
A proposed Novel Approach for Sentiment Analysis and Opinion Miningijujournal
as the people are being dependent on internet the requirement of user view analysis is increasing
exponentially. Customer posts their experience and opinion about the product policy and services. But,
because of the massive volume of reviews, customers can’t read all reviews. In order to solve this problem,
a lot of research is being carried out in Opinion Mining. In order to solve this problem, a lot of research is
being carried out in Opinion Mining. Through the Opinion Mining, we can know about contents of whole
product reviews, Blogs are websites that allow one or more individuals to write about things they want to
share with other The valuable data contained in posts from a large number of users across geographic,
demographic and cultural boundaries provide a rich data source not only for commercial exploitation but
also for psychological & sociopolitical research. This paper tries to demonstrate the plausibility of the idea
through our clustering and classifying opinion mining experiment on analysis of blog posts on recent
product policy and services reviews. We are proposing a Nobel approach for analyzing the Review for the
customer opinion
International Journal of Ubiquitous Computing (IJU)ijujournal
International Journal of Ubiquitous Computing (IJU) is a quarterly open access peer-reviewed journal that provides excellent international forum for sharing knowledge and results in theory, methodology and applications of ubiquitous computing. Current information age is witnessing a dramatic use of digital and electronic devices in the workplace and beyond. Ubiquitous Computing presents a rather arduous requirement of robustness, reliability and availability to the end user. Ubiquitous computing has received a significant and sustained research interest in terms of designing and deploying large scale and high performance computational applications in real life. The aim of the journal is to provide a platform to the researchers and practitioners from both academia as well as industry to meet and share cutting-edge development in the field.
USABILITY ENGINEERING OF GAMES: A COMPARATIVE ANALYSIS OF MEASURING EXCITEMEN...ijujournal
Usability engineering and usability testing are concepts that continue to evolve. Interesting research studies
and new ideas come up every now and then. This paper tests the hypothesis of using an EDA-based
physiological measurements as a usability testing tool by considering three measures; which are observers‟
opinions, self-reported data and EDA-based physiological sensor data. These data were analyzed
comparatively and statistically. It concludes by discussing the findings that has been obtained from those
subjective and objective measures, which partially supports the hypothesis.
SECURED SMART SYSTEM DESING IN PERVASIVE COMPUTING ENVIRONMENT USING VCSijujournal
Ubiquitous Computing uses mobile phones or tiny devices for application development with sensors
embedded in mobile phones. The information generated by these devices is a big task in collection and
storage. For further, the data transmission to the intended destination is delay tolerant. In this paper, we
made an attempt to propose a new security algorithm for providing security to Pervasive Computing
Environment (PCE) system using Public-key Encryption (PKE) algorithm, Biometric Security (BS)
algorithm and Visual Cryptography Scheme (VCS) algorithm. In the proposed PCE monitoring system it
automates various home appliances using VCS and also provides security against intrusion using Zigbee
IEEE 802.15.4 based Sensor Network, GSM and Wi-Fi networks are embedded through a standard Home
gateway.
PERFORMANCE COMPARISON OF ROUTING PROTOCOLS IN MOBILE AD HOC NETWORKSijujournal
Routing protocols have an important role in any Mobile Ad Hoc Network (MANET). Researchers have
elaborated several routing protocols that possess different performance levels. In this paper we give a
performance evaluation of AODV, DSR, DSDV, OLSR and DYMO routing protocols in Mobile Ad Hoc
Networks (MANETS) to determine the best in different scenarios. We analyse these MANET routing
protocols by using NS-2 simulator. We specify how the Number of Nodes parameter influences their
performance. In this study, performance is calculated in terms of Packet Delivery Ratio, Average End to
End Delay, Normalised Routing Load and Average Throughput.
The document compares the performance of various optical character recognition (OCR) tools. It analyzes eight OCR tools - Online OCR, Free Online OCR, OCR Convert, Convert image to text.net, Free OCR, i2OCR, Free OCR to Word Convert, and Google Docs. The document provides sample outputs of each tool processing the same input image. It then evaluates the tools based on character accuracy, character error rate, special symbol accuracy, and special symbol error rate to determine which tools most accurately convert images to editable text.
Optical Character Recognition (OCR) is a technique, used to convert scanned image into editable text
format. Many different types of Optical Character Recognition (OCR) tools are commercially available
today; it is a useful and popular method for different types of applications. OCR can predict the accurate
result depends on text pre-processing and segmentation algorithms. Image quality is one of the most
important factors that improve quality of recognition in performing OCR tools. Images can be processed
independently (.png, .jpg, and .gif files) or in multi-page PDF documents (.pdf). The primary objective of
this work is to provide the overview of various Optical Character Recognition (OCR) tools and analyses of
their performance by applying the two factors of OCR tool performance i.e. accuracy and error rate.
DETERMINING THE NETWORK THROUGHPUT AND FLOW RATE USING GSR AND AAL2Rijujournal
In multi-radio wireless mesh networks, one node is eligible to transmit packets over multiple channels to
different destination nodes simultaneously. This feature of multi-radio wireless mesh network makes high
throughput for the network and increase the chance for multi path routing. This is because the multiple
channel availability for transmission decreases the probability of the most elegant problem called as
interference problem which is either of interflow and intraflow type. For avoiding the problem like
interference and maintaining the constant network performance or increasing the performance the WMN
need to consider the packet aggregation and packet forwarding. Packet aggregation is process of collecting
several packets ready for transmission and sending them to the intended recipient through the channel,
while the packet forwarding holds the hop-by-hop routing. But choosing the correct path among different
available multiple paths is most the important factor in the both case for a routing algorithm. Hence the
most challenging factor is to determine a forwarding strategy which will provide the schedule for each
node for transmission within the channel. In this research work we have tried to implement two forwarding
strategies for the multi path multi radio WMN as the approximate solution for the above said problem. We
have implemented Global State Routing (GSR) which will consider the packet forwarding concept and
Aggregation Aware Layer 2 Routing (AAL2R) which considers the both concept i.e. both packet forwarding
and packet aggregation. After the successful implementation the network performance has been measured
by means of simulation study.
Building a Raspberry Pi Robot with Dot NET 8, Blazor and SignalR - Slides Onl...Peter Gallagher
In this session delivered at Leeds IoT, I talk about how you can control a 3D printed Robot Arm with a Raspberry Pi, .NET 8, Blazor and SignalR.
I also show how you can use a Unity app on an Meta Quest 3 to control the arm VR too.
You can find the GitHub repo and workshop instructions here;
https://bit.ly/dotnetrobotgithub
Google Calendar is a versatile tool that allows users to manage their schedules and events effectively. With Google Calendar, you can create and organize calendars, set reminders for important events, and share your calendars with others. It also provides features like creating events, inviting attendees, and accessing your calendar from mobile devices. Additionally, Google Calendar allows you to embed calendars in websites or platforms like SlideShare, making it easier for others to view and interact with your schedules.
DTWDIR: AN ENHANCED DTW ALGORITHM FOR AUTISTIC CHILD BEHAVIOUR MONITORING
1. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
DOI:10.5121/iju.2016.7201 1
DTWDIR: AN ENHANCED DTW ALGORITHM FOR
AUTISTIC CHILD BEHAVIOUR MONITORING
Salwa O Slim1
, Ayman Atia1
, and Mostafa-Sami M.Mostafa1
1
HCI-LAB, Department of Computer Science, Faculty of Computers and Information,
Helwan University, Cairo, Egypt
ABSTRACT
Autism has symptoms can hardly be recognized in the early stages of the disease, and it affects the child's
mental health on the long term. Autism can be identified by parents monitoring to the child and diagnosed
by psychiatrists using an international standard checklist. The checklist questions should be answered by
the parent and psychiatrist to determine the risk level of autism (high, medium, or low risk). It is hard for
parents to monitor more than 20 child's behaviours at the same time regardless lack of accuracy for
answering on most of these questions. We propose a system for monitoring autistic child behaviours by
analysing accelerometer data collected from wearable mobile device. The behaviours are recognized by
using a novel algorithm called DTWDir that based on calculating displacement and direction between two
signals. DTWDir is evaluated by comparing it to KNN, classical Dynamic Time Warping (DTW), and One
Dollar Recognition ($1) algorithms. The results show that DTWDir accuracy is higher than the others.
KEYWORDS
Autistic child, behaviours monitoring, DTW, KNN, One Dollar recognition
1. INTRODUCTION
According to a new report from the U.S. Centres for Disease Control and Prevention, there is a
30% increase in autistic children ratio than two years ago. Autism spectrum disorder (ASD) and
autism are both general terms for a group of complex disorders of brain development. These
disorders are characterized -in varying degrees- by difficulties in social interaction, verbal and
nonverbal communication, and repetitive behaviours. Psychiatrist who works with children must
deal with a great variety of behavioural and emotional problems therefore some of researches
provided standardized assessment and documentation of such problems and requires a little effort
by the psychiatrist. Those standardized assessments, may be a questionnaire [1] [2] or a checklist
standard [3] [4], are filled by parents and psychiatrist. The parents have a problem answering all
the questions accurately, because it requires parents' awareness to all the behaviours that result
from the autistic child and monitor the child for a long time attentively. Figure1 displays script
from standardized assessment Questions.
2. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
2
Figure 1. Script from modified checklist standard assessment for autism in toddlers [5]
The psychiatrists have to calculate the total score of the checklist questions and consequently
determining the level of autism risk. If the total score is 0-2 that means low risk, 3-7 is medium
risk, and 8-20 is high risk.
The majority of researches attempt to automatically identify the autistic child by monitoring
repetitive behaviours rather than social and communication defects [6][7] [8] which are easily
identified by using standardized assessments. This article is offering an automated methodology
to answer some of the standard checklist questions. Those questions are focused on
The ability of the child to copy what you do such as clapping like you does.
If the child has imagination and plays pretend or make-believe such as imaging he/she
drink from the cup.
If the child does stereotypical motor movement such as flapping hand.
And if the child interacts and communicates with the others such as acting goodbye to the
others.
It is hard to monitor and collect children's behaviours let alone autistic children, therefore we
have the challenge to collect data and apply an algorithm that can recognize the child's behaviour
using a small dataset. We collected the four behaviours (goodbye wave, drinking, hand flapping,
clapping) from a normal children by using accelerometer build in the smartphone and recognized
those behaviours using a novel algorithm called DTWDir that is based on comparing signals
displacement and direction. The displacement value is calculated by using DTW classifier and
direction value is calculated by collecting the direction variation -the difference in signs- between
two signals.
We evaluated our approach by comparing to KNN, One Dollar Recognition ($1), and classical
Dynamic Time Warping (DTW) classifiers. KNN accuracy is directly proportional the size of the
dataset even if the accuracy change is not effective [24]. Many researches depended on reducing
the size of datasets [25], [26], [27], and [28] when it applied KNN, as it has to use all the training
examples on each test which increases execution time with large datasets [29][30][31]. Bagnal et
al. [32] determined that 1-NN with an elastic measure such as DTW is the best approach for
smaller data sets, but that as the number of series increases "the accuracy of elastic measures
converges with that of Euclidean distance". Euclidian distance calculation is the square of the
difference between two points and that means the value's sign are lost therefore we developed
DTWDir which considers the absolute distance value and also its sign. By comparing DTWDir to
KNN, $1, and DTW the results showed that DTWDir accuracy is higher than the others and
reached to 93% when using three axes together.
3. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
3
The rest of the paper is as follows: we review related work in section2. Section3 explains the
DTWDir classifier. Dataset is explained in section 4. The experiments and results are explained in
section5. Section6 include the Discussions. Conclusion is explained in section7.
2. RELATED WORK
The related work is divided into two parts the first one is related to some of the classifiers that are
used for recognizing living activities and the second part displays the previous works which are
related to the autistic child.
2.1. Living Activities
Many researchers monitored the living activities by using smart phone accelerometer. Sang et al.
[9] and Ortiz et al. [10] have proposed a system to monitor some of living activities by using
KNN for recognition and they compared it with others classifiers. The authors found that KNN
accuracy reach to 74%, 88% in [9],[10] respectively. Sang et al. found that the recognition results
of Driving, sitting, and On-table activities reach to 100% but the results of Downstairs and
upstairs activities are poor due to the ambiguity of accelerometer and gyroscope data therefore the
authors suggests in future that those two activities should be concerned as a single activity for
better recognition results.
Paiyarom et al. [18] recognized ten activities of daily living by using DTW classifier. The data
are collected from two subjects – female, male- and they found that the mean accuracy of DTW is
91% even though the dataset training size is small. Muscillo et al. [19] compared between
decision tree C4.5 and DTW classifiers for recognizing low-level food preparation activities. The
data are collected from four sensors equipped kitchen utensils (knifes and spoon) about 6 hours.
They found that DTW accuracy is better than C4.5. Li et al. [20] developed a new algorithm
depends on using cubic spline interpolation and Derivative DTW by using Derivative DTW
classifier on the interpolated original data. They compared between DTW, Derivative DTW, and
the new algorithm on twelve different datasets and found that the accuracy of the new algorithm
is beneficial when DTW could not recognize the specific dataset and can produce much less
number of warping. Nayak et al. [21] proposed a new algorithm that is similar to the algorithm of
the Li et al. [20] but it used DTW classifier rather than Derivative DTW on the interpolated
original data. The authors used five different probabilistic distance measurements for DTW and
they evaluated the new algorithm on three different datasets; sign recognition (with large number
of possible classes), gesture recognition (with person variations), and classification of human
interaction sequences (with segmentation problems). The results showed that DTW without
reduction dimensions (original data) is better than DTW with reduction (interpolation).
2.2. Autistic Child
Most of the previous researches interest was to monitor stereotypical motor movements to
identify the autistic child and did not care about the interaction and communication of the child
with others such as Westeyn et al. [8] and Albinali et al. [6]. Westeyn et al. used 3-axes
accelerometers for detecting stereotypical motor movements from individuals mimicking the
actual behaviours. They achieved 69% of hand flapping behaviour, which automatically and
accurately is detected by using Hidden Markov Model, when Insertion errors are reduced.
Albinali et al. published two articles. Albinali et al. 2009 collected the data from 3-axes
accelerometer readings from 6 autistic children. They used J48 Decision Tree classifier, which is
an open source Java implementation of the decision tree algorithm(C4.5) in the WEKA data
mining tool, to identify stereotypical motor movements in the classroom and laboratory and they
4. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
4
found the mean accuracy is 88.6%, 89.5% respectively [7]. A major concern is the high false
positive rates averaging 0.08 across all participants. For intervention applications that target
specific stereotypical motor movements, the system would incorrectly deliver the intervention 8%
of the time when the participant is not engaged in the behaviour. Albinali et al. 2012 added expert
and non-experts system using software on mobile phones to annotate stereotypical motor
movements for classifier training [6]. They didn't collect any sensor data related to child's
interaction and communication.
Chuah et al. [11] developed Smartphone-Based Autism Social Alert (SASA) for automatically
detecting autistic child's stereotypical behaviours early and send alert to caregivers. They
recognized audio background with the occurrence of stereotypical behaviours so as to identify
potential environmental factors that may trigger such behaviours. Chuah et al. also used the same
algorithm (J48 Decision Tree classifier) of the Albinali et al 2009 and 2012 but Chuah et al.
distinguished stereotypical behaviours like foot tapping, jumping, hand waving, walking, sitting
with an accuracy of 85%.
Postawka et al. [22] used kinect sensor and the skeleton structure for monitoring the autistic child.
HMM algorithm is used for activity modelling and emotions recognition. They applied
experiments on standing and sitting activities. Gonçalves et al. [23] aimed to identify if the Kinect
sensor and the eZ430-Chronos watch with the accelerometers can be used as a time-efficient tool
to automatic detect child's behaviours. Two behaviours (hand flapping and body rocking
movement) are collected from four autistic children by using Kinect sensor and wireless
accelerometers placed on each wrist and chest. They used DTW recognizer and found that DTW
accuracy when using accelerometers sensors (76%) is better than when using kinect sensor
(51%).
3. DTWDIR CLASSIFIER
The DTWDir is an improvement to DTW classifier. DTW is a technique for matching two time
series' data [15]. The main purpose is to align two series in an optimal way, i.e. minimizing costs
for the warping path [16]. This path can be found very efficiently using dynamic programming to
evaluate the following recurrence which defines the cumulative distance ( )as the Euclidian
distance found in the current cell and the minimum of the cumulative distances of the
adjacent elements:
( ) ( ) ( ) ( )
( ) ( ) .
DTWDir depends on two main measured values on the selected Window length. These values are
the displacement value which is calculated by DTW classifier and the direction value which aim
to find the sign variation between the two series.
The following example illustrates the different between KNN, DTW, and DTWDir classifiers on
three synthetic signals (Temp1, Temp2, Query) for X-axis, Y-axis in Figure 2.
We can classify Query as Temp1 series because the directions of both signals in X-axis are
(positive (+Ve) then negative (-Ne)) but in Temp2 are (-Ne) then (+Ve). In the Y-Axis three
signals are the same.
5. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
5
Figure 2. Three synthetic signals in (X,Y) axes
In KNN, Query signal is classified as Temp1 or Temp2 because the displacement of Query to
Temp1 and Temp2 are the same. DTW classified Query as Temp2 series because it measured the
displacement between Query and Temp2 (≈15.87) is less than the displacement between Query
and Temp1 (≈17.25).DTWDir was able to classify Query as Temp1 series because its number of
sign variation is smaller than Temp2.
DTWDir Pseudo Code
Given Testing Template that has sequence of points , where
Given Training Templates with labels respectively and each
training template has sequence of points , where is number of training
templates and
Given Window Length (with seconds)
Given Overlap (with seconds)
Output where is the closet template for the
//preparing the testing template by Selecting a period of time equal to WL in the center of
period
̀ , where ⁄
⁄
//comparing each training template with testing template ̀
For i=1 to M Do Loop
//comparing one training template with ̀
For j=1 to Do Loop
̀
.pushback(DTWAlgorithm( ̀ , ̀ ), Direction( ̀ , ̀ ))
j=j+SOL
END Loop
Ranking with direction ascending order
TrainingResultsArrr.pushback(
End Loop
Ranking TrainingResultsArr with Distance ascending order
= Label of (TrainingResultsArr [0])
4. DATA SET
Our data set is collected by 'Lenovo' smart phone sensor which its sampling rate is approximately
98 samples per second. We recorded 6 sec for each try of four behaviours (goodbye, hand
-5
0
5
10
0 1 2 3 4
Temp1 Query Temp2
+Ve -Ne
X-Axis
0
10
20
0 1 2 3 4
Temp2 Query Temp1
Y-Axis
6. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
6
flapping, drinking, and clapping) from 3 children with ages 4±1. We collected 10 tries per
behaviour per child. The acceleration and orientation values are collected for each try per
behaviour.
5. EXPERIMENTS AND RESULTS
5.1 Experiments Setup
A C++ application was developed to compare DTWDir with KNN, One Dollar recognizer, and
DTW classifiers. K-fold cross validation with K=3 is used to split data into 3 parts that each part
is one child but before any manipulation we used high-pass filter to remove gravity from the
measured acceleration [17]. As the frequency of gravity is low, it is considered that filtering low
frequency acceleration is effective in reducing the impact of gravity. There are many criteria
effects on DTWDir quality that are in the following list.
Sensor type: we recorded the acceleration and the orientation data.
Pre-processing operations: we have 3 ways for entering the data to a novel algorithm;
o Raw data: without any selection or changing in the data. This attribute isn't
available to apply on KNN algorithm, because all training and testing templates
should have the same length.
o Interpolation: it is applied on the raw data to ensure all templates have the same
length [12].
o Selected 5 sec: it's like raw data but to guarantee the number of sampling in all
templates are equal, we selected 5 sec from the originally recorded 6 sec of raw
data.
Axis type: there are three axes. We need to know which axis is better than the others.
As we mentioned before we aim to measure the accuracy of DTWDir by comparing with three
others classifiers but there are many criteria affects on our algorithm therefore we are going to
apply changes on those criteria and observe how the DTWDir will react. We summarized
experiments description and goals in Table 1.
Table 1: summarized experiments analysis
Exp Description Goals
#1 Comparison between KNN,
DTW, $1, and DTWDir using
the values of the sensor's axis
X, Y, and Z individually and
the three axes together using
two types of sensors
acceleration and orientation.
1. Measure the accuracy of each algorithm
when sensor types are Orientation (OR) or
Acceleration (AC).
2. Identify the best sensor for each algorithm.
3. Determine which axis is better than the
others in AC and OR sensors.
#2 Comparison between four
algorithms when pre-
processing operations are
changed.
1. Measure the accuracy of each algorithm
when the pre-processing operation is
changed
2. Determine which pre-processing operation is
the best for increasing the algorithms
accuracy.
7. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
7
#3 Comparison between four
algorithms when we selected 5
sec from raw data and used
both the three axes together
and the Z axis alone.
1. Measure the accuracy of the behaviours in
each algorithm.
2. Determine the best algorithm is able to
detect hand flapping behaviour.
5.2 Results
In Experiment #1, we interpolate the raw data to 300 samples in the pre-processing operation and
put the value of overlap with 0.5 sec and window length with 2 sec. The results are shown on
Table 2.
Table 2: The comparison results accuracy of four classifiers according to two sensors types (OR and AC)
for each axis individually and three axes together
Sensor
Type
Classifier
OR AC
X Y Z X,Y,Z AVG X Y Z X,Y,Z AVG
KNN 69 74 62 80 71.25 63 60 54.6 58 58.9
$1 39.6 37.3 59 49.6 46.3 36.3 47.3 35.3 59.6 44.6
DTW 80.6 84.6 77 90 83 64 83 59.6 90.6 74.3
DTWDir 83.3 88.6 86 89 86.7 82 87.3 64 90.6 80.9
68.1 71.1 71 77.1 61.3 69.4 53.3 74.9
In Experiment #2, we used raw data in pre-processing step, accelerometer sensor type, and X-
axis. Because raw data doesn't have the same length in all tries, KNN couldn't apply on raw data
attributes. The results are shown on the Table 3.
Table 3: The accuracy of four classifiers when pre-processing operation is changed
Entering data
Classifier
Raw data Interpolation
N=300
Selected 5 secs
KNN 63% 60.6%
DTW 76% 64% 75.6%
DTWDir 80% 82.3% 84.6%
$1 34.6% 36.6% 37.6%
Average 63.5% 61.4% 64.6%
In Experiment #3, we used Acceleration sensor type, overlap value is 0.25 sec and window length
value is 2 sec. the results are shown on the Table 4.
8. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
8
Table 4: Accuracy of each behaviour
6. DISCUSSIONS
We applied KNN [9] on our dataset and we found that DTWDir is better than KNN as displayed
in Table 2. DTWDir is higher than KNN in all axis in both acceleration and orientation sensors.
Table 2 displayed that all algorithms accuracy in the orientation data sensor is better than the
acceleration data. It's better to use three axes together rather than single axis for all algorithms in
both acceleration and orientation sensors. The accuracy of DTWDir in both sensor types is higher
than the others.
The interpolation pre-processing may change the original data values (raw data) but it keeps the
form of the signal pattern [13] therefore DTW accuracy in interpolation is less than in raw data
attributes as in Table 3 because the purpose of DTW algorithm matches the best wrapping path
between the two signals [14] but the DTWDir accuracy was not affected. DTW accuracy in 5 sec
is better than in interpolation but it still was not higher than raw data pre-processing. The authors
[20] [21] developed new algorithms depend on interpolation pre-processing and DTW classifier.
The article [20] displayed that the new algorithm accuracy was not beneficial except in the case
of DTW failure to recognize the behaviour and also [21] showed that DTW with raw data is better
than DTW with interpolation pre-processing. The selected 5 sec pre-processing is better than the
others in all algorithms except for KNN because it depends on Euclidian distance in calculation.
For sure the speed of algorithms in the interpolation is higher than the others pre-processing
operations because the length of data signals in the interpolation is 300 points and in the selection
5 sec pre-processing operation is 490 points (5*98) where 98 is the sampling rate. DTWDir is
better than the other algorithms in all pre-processing operations.
Most of previous researches [6] [7] [11] used decision tree algorithm (C4.5) for recognizing
stereotypical motor movements to identify the autistic child but Muscillo et al. [19] displayed that
DTW is better than C4.5 when they compared between them. As mentioned before, stereotypical
motor movements are not enough to identify the autistic child therefore they were not in
consideration, however we compared hand flapping detection accuracy in both cases when
DTWDir achieved highest, and lowest accuracy using three axes together, and Z-Axis
respectively. Table2 displayed that the accuracy of algorithms in three axes together is better than
the others but Z-axis is the least one. Table 4 showed that DTWDir is better than KNN, $1 and
DTW in both cases. Westeyn et al. [8] achieved 69% by using Hidden Markov Model algorithm
and a 3 axes accelerometer, but DTWDir was able to detect hand flapping behaviour with
percentage 73.3%.
Classifiers
Behaviours
KNN $1 DTW DTWDir
3 Axes Z-Axis 3 Axes Z-
Axis
3 Axes Z-Axis 3 Axes Z-Axis
Clapping 100% 100% 46.6% 36.6% 100% 83.3% 100% 86.6%
Drinking 100% 100% 90% 93.3% 100% 73.3% 100% 96.3%
Flapping 20% 20% 33.3% 13.3% 66.6% 3.3% 73.3% 33.3%
Goodbye 23.3% 23.3% 13.3% 23.3% 100% 80% 100% 73.3%
Average 61% 61% 46% 42% 91.6% 60% 93% 72%
9. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
9
7. CONCLUSION
Psychiatrists are interested in the child's interaction and communication with the others not only
stereotypical motor movement behaviours. They used checklist standard assessment to determine
the risk level of autism. The checklist contains several yes or no questions each question aims to
measure some of child's behaviour such as (interaction with the others, communication skills, and
stereotypical motor behaviours). Those questions have to be answered by parents and
psychiatrists but it is hard for the parents to monitor the child for a long time. Psychiatrists do a
lot of effort to diagnose the children as they have a variety of behavioural and emotional
problems. This paper will help them by providing an automated technique to answer some of the
checklist questions. The questions in consideration are answered by evaluating child's behaviours
which are monitored by accelerometer. We collected four behaviours; three behaviours are
dealing with interaction and communication (goodbye, drinking, clapping), one behaviour is
dealing with stereotypical movement (hand flapping). The dataset is recorded from three child's
behaviours that mimic the autistic child. We applied KNN, $1 recognition, and DTW classifiers
on our dataset but we found that the average accuracy of three algorithms are 71%, 46%, 83%
respectively using orientation sensor type and 58.9%, 44.6%, 74% respectively using acceleration
sensor type. DTW is better than $1 and KNN in both acceleration and orientation sensor types
therefore we attempt to improve the accuracy of DTW recognition by developing a novel
algorithm based on DTW that is called DTWDir. We evaluated the novel algorithm by comparing
it to KNN, $1 and DTW on our dataset. From the previous experiments we found that DTWDir is
better than the other algorithms but it consumes high execution time. In the future work we will
try to reduce the execution time by combining KNN with DTWDir.
References
[1] M. Posserud, A. Lundervold and C. Gillberg, "Autistic features in a total population of 7-9-year-
old children assessed by the ASSQ (Autism Spectrum Screening Questionnaire)", J Child Psychol
& Psychiat, vol. 47, no. 2, pp. 167-175, 2006.
[2] M. Mattila, K. Jussila, S. Linna, M. Kielinen, R. Bloigu, S. Kuusikko-Gauffin, L. Joskitt, H.
Ebeling, T. Hurtig and I. Moilanen, "Validation of the Finnish Autism Spectrum Screening
Questionnaire (ASSQ) for Clinical Settings and Total Population Screening", J Autism Dev
Disord, vol. 42, no. 10, pp. 2162-2180, 2012.
[3] Validation of the Modified Checklist for Autism in Toddlers, Revised With Follow-up (M-CHAT-
R/F)", PEDIATRICS, vol. 133, no. 1, pp. X5-X5, 2013.
[4] Sunita and J. Bilszta, "Early identification of autism: A comparison of the Checklist for Autism in
Toddlers and the Modified Checklist for Autism in Toddlers", Journal of Paediatrics and Child
Health, vol. 49, no. 6, pp. 438-444, 2012.
[5] J. M. Kleinman et al., "The modified checklist for autism in toddlers: A follow-up study
investigating the early detection of autism spectrum disorders," Journal of Autism and
Developmental Disorders, vol. 38, no. 5, pp. 827–839, Sep. 2008.
[6] F. Albinali, M. Goodwin and S. Intille, "Detecting stereotypical motor movements in the
classroom using accelerometry and pattern recognition algorithms", Pervasive and Mobile
Computing, vol. 8, no. 1, pp. 103-114, 2012.
[7] Fahd Albinali, Matthew S. Goodwin, and Stephen S. Intille. "Recognizing stereotypical motor
movements in the laboratory and classroom: a case study with children on the autism spectrum",
In Proceedings of the 11th international conference on Ubiquitous computing (UbiComp '09).
ACM, New York, pp.71-80, 2009.
[8] T. Westeyn, K. Vadas, X. Bian, T. Starner and G. D. Abowd, "Recognizing mimicked autistic
self-stimulatory behaviours using HMMs," Wearable Computers, 2005. Proceedings. Ninth IEEE
International Symposium on, pp. 164-167, 2005.
10. International Journal of UbiComp (IJU), Vol.7, No.2, April 2016
10
[9] V. Sang, N. Thang, V. Van Toi, N. Hoang and T. Khoa, "Human Activity Recognition and
Monitoring Using Smartphones", IFMBE Proceedings, vol. 46, pp. 481-485, 2015.
[10] J. Ortiz, "Smartphone-based human activity recognition", Springer International Publishing, 2015.
[11] M. Chuah and M. Diblasio. "Smartphone Based Autism Social Alert System," Mobile Ad-hoc and
Sensor Networks (MSN), 2012 Eighth International Conference on, Chengdu, pp. 6-13, 2012.
[12] Wobbrock, J.O., Wilson, A.D. and Li, Y. "Gestures without libraries, toolkits or training: A $1
recognizer for user interface prototypes", Proceedings of the ACM Symposium on User Interface
Software and Technology (UIST '07). Newport, Rhode Island New York: ACM Press, pp. 159-
168, 2007.
[13] G. Opfer, "Gary D. Knott, Interpolating Cubic Splines", Journal of Approximation Theory, vol.
112, no. 2, pp. 319-321, 2001.
[14] H. Ding, G. Trajcevski, P. Scheuermann, X. Wang and E. Keogh, "Querying and mining of time
series data", Proc. VLDB Endow., vol. 1, no. 2, pp. 1542-1552, 2008.
[15] C. Myers and L. Rabiner, "A Comparative Study of Several Dynamic Time-Warping Algorithms
for Connected-Word Recognition", Bell System Technical Journal, vol. 60, no. 7, pp. 1389-1409,
1981.
[16] M. Eamonn J. Keogh, "Derivative Dynamic Time Warping", In First SIAM International
Conference on Data Mining (SDM’2001), 2001.
[17] D. Goehl and D. Sachs. "Motion sensors gaining inertia with popular consumer electronics". White
Paper, IvenSense Inc, 2007.
[18] S. Paiyarom, P. Tungamchit, R. Keinprasit and P. Kayasith, "Activity monitoring system using
Dynamic Time Warping for the elderly and disabled people", 2009 2nd International Conference
on Computer, Control and Communication, pp. 1-4, 2009.
[19] R. Muscillo, S. Conforto, M. Schmid, P. Caselli and T. D'Alessio, "Classification of Motor
Activities through Derivative Dynamic Time Warping applied on Accelerometer Data", 2007 29th
Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp.
4930-4933, 2007.
[20] H. Li, X. Wan, Y. Liang and S. Gao, "Dynamic Time Warping Based on Cubic Spline
Interpolation for Time Series Data Mining," Data Mining Workshop (ICDMW), 2014 IEEE
International Conference on, Shenzhen, pp. 19-26, 2014.
[21] S. Nayak, S. Sarkar and B. Loeding, "Distribution-Based Dimensionality Reduction Applied to
Articulated Motion Recognition", IEEE Transactions on Pattern Analysis and Machine
Intelligence, vol. 31, no. 5, pp. 795-810, 2009.
[22] A. Postawka and P. Śliwiński, "Recognition and Modeling of Atypical Children Behavior",
Artificial Intelligence and Soft Computing, vol. 9119, pp. 757-767, 2015.
[23] N. Gonçalves, J. L. Rodrigues, S. Costa and F. Soares, "Automatic detection of stereotyped hand
flapping movements: Two different approaches," RO-MAN, 2012 IEEE, Paris, pp. 392-397, 2012.
[24] S. Yu, S. Guo and S. Publishing, "Big Data Concepts, Theories, and Applications | Shui Yu |
Springer", Springer.com, 2016. [Online]. Available:
http://www.springer.com/la/book/9783319277615.
[25] P. Hart, "The Condensed Nearest Neighbour Rule," IEEE Transactions on Information Theory,
vol. 14, pp. 515-516, 1968.
[26] G. Gates, "The Reduced Nearest Neighbour Rule," IEEE Transactions on Information Theory, vol.
18, pp. 431-433, 1972.
[27] M. Kubat and M. Jr, "Voting Nearest-Neighbour Subclassifiers," in Proceedings of the 17th
International Conference on Machine Learning, ICML-2000 , Stanford, CA, pp. 503-510, 2000.
[28] D. R. Wilson and T. R. Martinez, "Reduction Techniques for Exemplar-Based Learning
Algorithms," Machine learning, vol. 38, no. 3, pp. 257-286, 2000.
[29] G. Guo, H. Wang, D. Bell, Y. Bi, and K. Greer, "KNN Model-Based Approach in Classification,"
Lecture Notes in Computer Science, vol. 2888, pp. 986-996, 2003.
[30] N. Bhatia and A. Vandana,"Survey of Nearest Neighbor Techniques", (IJCSIS) International
Journal of Computer Science and Information Security , vol. 8, no. 2, pp. 302-305, 2010.
[31] A. Basheer, M. Ali, G. Awad, and A. Ali, "Solving the Problem of the K Parameter in the KNN
Classifier Using an Ensemble Learning Approach", ( IJCSIS) International Journal of Computer
Science and Information Security, Vol. 12, No. 8, August 2014 .
[32] A. Bagnall, L. Davis, J. Hills, and J. Lines, "Transformation Based Ensembles for Time Series
Classification", Proceedings of the 12th (SIAM) International Conference on Data Mining, pp.
307—319, 2012.