06/13/2025
PAPER ID: 1009
PAPER TITLE:
STRESS DETECTION SYSTEM USING
DECISION TREE INTEGRATING
WEARABLES SENSORS
Presented By:
Ankit Thakur
Galgotias University
1
International Conference on Artificial
Intelligence and Sustainable Innovation 2025
(ICAISI-2025)
06/13/2025
OUTLINE
• Abstract
• Introduction
• Related Work
• Comparison with Existing Work
• Methodology
• Results and Discussion
• Conclusion and Future Work
• Reference
2
06/13/2025
ABSTRACT
Anxiety is a natural part of life and can be caused by physical, psychological, and social factors. It can
help control mental health and moderate physical stress, but it can also lead to more serious conditions
like depression, anxiety, and cardiovascular stress. It should be identified early. Since mental stress is
taboo in our society, this paper aimed to develop a stress detection system based on bio-signals captured
from the body using wearable devices with physiological and motion sensors and machine learning
techniques. Stress not only affects a person's mental health but also their physical health and
surroundings. The three-dimensional axis study also analysis [EMG], [AB] promising utilized [LDA],
and other random electrodermal addition accuracies machine forest activity to and of [EDA] are just a
few of the physiological characteristics that these devices may monitor and record. We know that there
are numerous factors which are very significant to capture like heartbeat sensors [ECG], Galvanic Skin
Response [GSR], facial expressions & SPO2 level. Our primary goal is to develop a hardware and
software integrated system that can transform the medical industry by tracking or curing mental
stress in a low-cost manner. With the aid of wearables, the user enters all the crucial information into
the software, and the results are encouraging as well. The user can record many data points to evaluate
progress, and the software will provide the user's current mental health state. The accuracy of the
software is between 85 and 90 percent and can be improved further.
3
06/13/2025
INTRODUCTION
Stress has become a key factor in contemporary living. Stress is a multidimensional
emotional and physical reaction to imbalance between outside stressors and inside coping
skills (Banerjee et al. 2023). According to the American Institute of Stress [2020], stress
levels in the workplace is extremely high, with 80% of workers invested feeling stressed at
their jobs and 42% of employees reporting that their coworkers need help managing stress .
Chronic stress can also lead to disorders such as cardiac arrhythmia, hypertension, and even
depression. Therefore, an important field of research has become detecting and alleviating
stress at an early stage. Traditional methods are based on psychological surveys and self-
reported data, which are subjective and prone to inaccuracies. In comparison, wearables and
biometric monitoring offer objective and continuous results. Furthermore, targeted
interventions, such as digital meditation applications, have been shown to booster self-
regulation capabilities through improved neurocognitive processes. For instance,
adolescents who engaged in closed-loop digital interventions exhibited enhanced attention
and reduced behavioral issues, underscoring the potential of these technologies to facilitate
stronger emotional regulation found in daily life. Thus, integrating such tools into mental
health practices holds promise for improving individual well-being. 4
06/13/2025
RELATED WORK
5
The integration of machine learning for stress detection has gained momentum in recent years. Investigated
by author and help us to learn About "Wearable Sensor-Based Stress Detection Using Deep Learning
(Gedam et al. 2021). A Survey which includes physiological and motion signals, to identify stress levels.
Their model utilized features like ECG, respiration, and EDA, achieving a classification accuracy of
80.34% for three-class problems and 93.12% for binary classifications Similarly, Author tried to Overcome
past hurdles by explaining the Role of Skin Conductance in mental health, (Kuttala et al. 2023).
"Evaluation and Classification of Physical and Psychological Stress in Firefighters Using Heart Rate
variability measured physiological parameters such as skin conductance and respiration to detect stress
using Principal Component Analysis [PCA] and Linear Discriminant Analysis [LDA] (Ninh et al. 2022).
Their system achieved 80% accuracy new datasets like SWELL-KW further advanced research by
incorporating facial expressions, posture analysis, and computer logging data. In this paper author Used
machine learning to revolutionize the field of Mental stress with the help of decision Tree, (Öztekin et al.
2025) "Stress Detection Using Context-Aware Sensor Fusion from Wearable Devices” validated these
datasets and demonstrated their utility in classifying stress and non-stress conditions (Rashid et al. 2023).
This study builds upon these foundations by integrating multiple machines learning models and enhancing
their performance through deep learning approaches, making the system more robust and scalable
06/13/2025
TABLE 1 - COMPARISON WITH
EXISTING WORK
Reference Methods Objective
Rashid et al. 2022 Wearable Devices Stress Detection using Context Aware Sensor Fusion.
Ninh et al. 2022 Wearable Devices Independent Stress Detection Model.
Xu et al. 2024 SELF-CARE Detect Stress Using Physiological Signals
Rashid et al. 2022 SELF-CARE Aware Low-Power Edge Computing for Stress Detection
Vigouroux et al. 2025 Machine Learning Heart RateVariability-Based Mental Stress Detection
Kuttala et al. 2023 Multimodal CNN Feature Fusion for Stress Detection
6
06/13/2025
METHODOLOGY
In this project we are implementing both software and hardware sides so we can increase the
efficiency of our system we are going to collect important data like heartbeat, SPO2, GSR with the
help of hardware like wearables and going to put the values in our software to track mental health
which is going to increase overall efficiency.
1. Data Collection
Wearable sensors capable of tracking the following physiological data were used to collect data:
Movement patterns: 3-axis acceleration [ACC]. Electrocardiogram [ECG] to track heart activity.
Blood Volume Pulse [BVP] to monitor changes in blood flow. BODY TEMP [TEMP] to capture
changes in stress responses. Respiration [RESP] to quantify variations in the rate of breathing.
Electromyogram [EMG] to monitor muscle activity. Electrodermal Activity [EDA], which when it
comes to sweat gland activation, how much conductivity we get out of skin and shown in Fig.1.
7
06/13/2025
2. Feature Extraction
Several features contributing to detect mental health stress are physiological features,
Behavioural features, contextual features they help us to detect many important parameters
Like Respiration Rate, Heartrate, Electrodermal activity and other parameters like sleep
pattern, speech features facial expression and less evolving parameters like social interaction
and environmental factors (Ali et al. 2021). Preprocessing was performed on the collected
data for noise removal and value normalization. Input vectors were constructed by extracting
statistical features such as mean, variance, and peak-to-peak differences for each imaging
channel in each time series interval (Öztekin et al. 2025).
3. Machine Learning Models
The algorithms used for classification consisted of: K-Nearest Neighbour [KNN]: Useful
for non-linear data classification linear discriminant analysis [LDA]: For dimensional
reduction and class separation. Decision tree [DT]: For 0/1 classification, easy to
understand. AdaBoost [AB] — Boosting weak classifiers.
y=wTx (1)
Note - That the equation is used for linear discriminant analysis it is use to simplify the
high dimensional data into lower dimensional data.
8
06/13/2025
4. Deep Learning Approach
We employed Artificial Neural Networks [ANN] to model complex interdependencies in the data and obtained almost perfect
classification of stress and non-stress conditions (Srinivasan et al. 2023). Cognitive Symptoms: Constant Depression, Harmful
thoughts, ADHD Symptoms, OCD Symptoms, Inability to focus, Financial Problem, Being pessimistic. Physical Symptoms:
Constipation or diarrhea, frequent colds and flu, aches and pains, nausea, and dizziness discomfort in the chest, Fast heartbeat, loose
intestine, choking sensation, teeth grinding, frequent and urgent urination, fatigue, and weight loss or gain (Srivastava et al. 2019).
Emotional Symptoms: Tension, irritability or anger, restlessness, worries, difficulty relaxing, depression, general dissatisfaction,
anxiety and agitation, and moodiness are examples of emotional symptoms (Trivedi et al. 2022).
Figure 1. - Process execution Conductance [Source: Researchgate.net]
In figure 1. - Discussed machine learning models which includes BVP Which Measures Blood Volume Pulse, EDA Use for Measure
Skin Conductance as Well as Skin temperature And Skin.
9
06/13/2025
5. SYSTEM DESIGN
As we already discuss that this project work on both aspects' software and hardware:
Dashboard
It is the user interface provides every information about the website the things it contains.
Log In
Allows users to create an account within the system.
Visualization
Provides visual representations of the user's mental health data, allowing them to track their
metrics over time. (Shanmugasundaram et al. 2019).
Wearables
This includes smart watches, health bands, medical equipment Like Blood pressure machine
etc.
10
06/13/2025
Physiology Sensors
ECG, GSR, PPG, SPO2, EMG And Temperature Sensor.
• Behavioral Sensors
Accelometer, Gyroscope, Camera Module, Arduino. Microphone, Wi-Fi Module, Battery pack etc.
Figure 2. - No. of student vs. Type of Sensor
In figure 2. - as you can see that it shows the number of students involved in to track their mental health using wearables which
can record all the important vitals which are shown above It also shows the unavailability of these services in many areas.
11
EDA ECG EEG PPG EMG fNIRS RESP SKT MOVE PUPIL SpO2 BP
0
5
10
15
20
Multimodal Stuides
Unimodal Stuides
Total Studies
Type of sensor
Number
of
Student
06/13/2025
Figure 3. – System Flow Chart
12
In figure 3. - Discussed about important vitals that are needed to track mental health. which includes several sensors
like GSR, SPO2, ECG with the help of machine learning using decision tree.
06/13/2025
RESULTS AND DISCUSSION
As shown in Fig. 4, this project works on hardware and software aspect. In hardware aspect we use sensors
like - Fitbit, Apple Watch, GSS Sensors, Face detection to monitor user data which is going to help us to
determine the Stress level where as in software aspect we don’t use or own any hardware we just input all
the important data in the software and calculate the stress level of user but the problem is accuracy in
software aspect we have to comprise with accuracy but it is an cheap alternative so for now we mainly
depend on the software aspect of this project and in future we can also integrate hardware to increase the
accuracy. As for now, in this era of new technologies people are concerned about their mental health, but
they still don’t want to spend money on this type of hardware equipment so let’s continue with the software
Aspect. These results punctuate the trustability of integrating wearable detectors with machine literacy
algorithms for real- time stress monitoring. Compared to former styles that reckoned solely on single
physiological signals. This System Demonstrates Bettered Robustness by integrating Multiple Futures
Contemporaneously. These results punctuate the trustability of integrating wearable detectors with machine
literacy algorithms for real- time stress monitoring.
13
06/13/2025 14
Figure 4. - User Interface
In figure 4. - In this page the User Supposed to Fill his/her important vital which he/her
record with the help of wearables sensors. After entering all the vitals accurately. It shows you
the result whereas this page records all the previous data of the user so the user can record his
data which help him/her in overall well being.
06/13/2025
CONCLUSION AND FUTURE WORK
The Our system collects real-time physiological data from the user using various sensors that measure: Heart Rate &
Blood Pressure – Indicators of stress, anxiety, or calmness. This data is then processed by our software, which analyses
patterns and provides insights into the person’s mental state. Currently, our system achieves an accuracy rate of around
80-95%, and we are working on making it even more precise.
This Paper presents a scalable and accurate stress detection system leveraging wearable sensor data and Integrated
machine learning to formulate new ideas to control stress. The system surpasses existing models in accuracy and
flexibility, making it suitable for real-world applications such as workplace stress management and health monitoring.
Future research can focus on integrating facial recognition and speech analysis to complement physiological data,
enhancing the accuracy further.
Additionally, deploying this system in real-time smart phones software can enable broader adoption and continuous
monitoring system in real-time mobile applications can enable broader adoption and continuous monitoringenhancing the
accuracy further. Additionally, deploying this system in real-time mobile software can enable broader adoption and
continuous monitoring. Mental health tracking should not be limited to clinical settings or expensive devices.
Our goal is to make this technology affordable and accessible so that more people can monitor and manage their
mental health proactively. By bridging the gap between biometric data and mental well-being, we hope to
contribute to a world where mental health care is more effective, personalized, and within everyone’s reach.
15
06/13/2025
REFERENCE
1. Banerjee, J.S., Mahmud, M., & Brown, D. 2023. Heart rate variability-based mental stress detection: an explainable machine learning
approach. SN Computer Science 4(2): 176.
2. Gedam, S., & Paul, S. 2021. A review on mental stress detection using wearable sensors and machine learning techniques. IEEE Access 9:
84045-84066.
3. Kuttala, R., Subramanian, R., & Oruganti, V.R.M. 2023. Multimodal hierarchical CNN feature fusion for stress detection. IEEE Access 11:
6867-6878.
4. Le Vigouroux, S., Chevrier, B., Montalescot, L., & Charbonnier, E. 2025. Post-pandemic student mental health and coping strategies: A
time trajectory study. Journal of Affective Disorders.
5. M.S.F. Ali, M.A. Khan, & M.H. Rehmani. 2021. Wearable Sensor-Based Stress Detection Using Deep Learning: A Survey. IEEE Access 9:
12423-12441.
16
06/13/2025
REFERENCE
6. Ninh, V.T., Nguyen, M.D., Smyth, S., Tran, M.T., Healy, G., Nguyen, B.T., & Gurrin, C. 2022. An improved subject-independent stress
detection model applied to consumer-grade wearable devices. In International Conference on Industrial, Engineering and Other
Applications of Applied Intelligent Systems, pp. 907-919. Cham: Springer International Publishing.
7. Öztekin, G.G., Gómez-Salgado, J., & Yıldırım, M. 2025. Future anxiety, depression and stress among undergraduate students:
psychological flexibility and emotion regulation as mediators. Frontiers in Psychology 16: 1517441.
8. Rashid, N., Mortlock, T., & Al Faruque, M.A. 2022. Self-care: Selective fusion with context-aware low-power edge computing for stress
detection. In 2022 18th International Conference on Distributed Computing in Sensor Systems (DCOSS), pp. 49-52. IEEE.
9. Rashid, N., Mortlock, T., & Al Faruque, M.A. 2023. Stress detection using context-aware sensor fusion from wearable devices. IEEE
Internet of Things Journal 10(16): 14114-14127.
10. Shanmugasundaram, G., Yazhini, S., Hemapratha, E., & Nithya, S. 2019. A comprehensive review on stress detection techniques. In 2019
IEEE International Conference on System, Computation, Automation and Networking (ICSCAN), pp. 1-6. IEEE.
17
06/13/2025
REFERENCE
11. Trivedi, N.K., Tiwari, R.G., Witarsyah, D., Gautam, V., Misra, A., & Nugraha, R.A. 2022. Machine learning based
evaluations of stress, depression, and anxiety. In 2022 International Conference Advancement in Data Science, E-learning
and Information Systems (ICADEIS), pp. 1-5. IEEE.
12. Xu, G., Qin, R., Zheng, Z., & Shi, Y. 2024. An Adaptive System for Wearable Devices to Detect Stress Using Physiological
Signals. arXiv preprint arXiv:2407.15252.
18
06/13/2025
THANK YOU
19
06/13/2025
QUESTION AND ANSWERS?
20

updated_Paper ID - 1009 machine learning model

  • 1.
    06/13/2025 PAPER ID: 1009 PAPERTITLE: STRESS DETECTION SYSTEM USING DECISION TREE INTEGRATING WEARABLES SENSORS Presented By: Ankit Thakur Galgotias University 1 International Conference on Artificial Intelligence and Sustainable Innovation 2025 (ICAISI-2025)
  • 2.
    06/13/2025 OUTLINE • Abstract • Introduction •Related Work • Comparison with Existing Work • Methodology • Results and Discussion • Conclusion and Future Work • Reference 2
  • 3.
    06/13/2025 ABSTRACT Anxiety is anatural part of life and can be caused by physical, psychological, and social factors. It can help control mental health and moderate physical stress, but it can also lead to more serious conditions like depression, anxiety, and cardiovascular stress. It should be identified early. Since mental stress is taboo in our society, this paper aimed to develop a stress detection system based on bio-signals captured from the body using wearable devices with physiological and motion sensors and machine learning techniques. Stress not only affects a person's mental health but also their physical health and surroundings. The three-dimensional axis study also analysis [EMG], [AB] promising utilized [LDA], and other random electrodermal addition accuracies machine forest activity to and of [EDA] are just a few of the physiological characteristics that these devices may monitor and record. We know that there are numerous factors which are very significant to capture like heartbeat sensors [ECG], Galvanic Skin Response [GSR], facial expressions & SPO2 level. Our primary goal is to develop a hardware and software integrated system that can transform the medical industry by tracking or curing mental stress in a low-cost manner. With the aid of wearables, the user enters all the crucial information into the software, and the results are encouraging as well. The user can record many data points to evaluate progress, and the software will provide the user's current mental health state. The accuracy of the software is between 85 and 90 percent and can be improved further. 3
  • 4.
    06/13/2025 INTRODUCTION Stress has becomea key factor in contemporary living. Stress is a multidimensional emotional and physical reaction to imbalance between outside stressors and inside coping skills (Banerjee et al. 2023). According to the American Institute of Stress [2020], stress levels in the workplace is extremely high, with 80% of workers invested feeling stressed at their jobs and 42% of employees reporting that their coworkers need help managing stress . Chronic stress can also lead to disorders such as cardiac arrhythmia, hypertension, and even depression. Therefore, an important field of research has become detecting and alleviating stress at an early stage. Traditional methods are based on psychological surveys and self- reported data, which are subjective and prone to inaccuracies. In comparison, wearables and biometric monitoring offer objective and continuous results. Furthermore, targeted interventions, such as digital meditation applications, have been shown to booster self- regulation capabilities through improved neurocognitive processes. For instance, adolescents who engaged in closed-loop digital interventions exhibited enhanced attention and reduced behavioral issues, underscoring the potential of these technologies to facilitate stronger emotional regulation found in daily life. Thus, integrating such tools into mental health practices holds promise for improving individual well-being. 4
  • 5.
    06/13/2025 RELATED WORK 5 The integrationof machine learning for stress detection has gained momentum in recent years. Investigated by author and help us to learn About "Wearable Sensor-Based Stress Detection Using Deep Learning (Gedam et al. 2021). A Survey which includes physiological and motion signals, to identify stress levels. Their model utilized features like ECG, respiration, and EDA, achieving a classification accuracy of 80.34% for three-class problems and 93.12% for binary classifications Similarly, Author tried to Overcome past hurdles by explaining the Role of Skin Conductance in mental health, (Kuttala et al. 2023). "Evaluation and Classification of Physical and Psychological Stress in Firefighters Using Heart Rate variability measured physiological parameters such as skin conductance and respiration to detect stress using Principal Component Analysis [PCA] and Linear Discriminant Analysis [LDA] (Ninh et al. 2022). Their system achieved 80% accuracy new datasets like SWELL-KW further advanced research by incorporating facial expressions, posture analysis, and computer logging data. In this paper author Used machine learning to revolutionize the field of Mental stress with the help of decision Tree, (Öztekin et al. 2025) "Stress Detection Using Context-Aware Sensor Fusion from Wearable Devices” validated these datasets and demonstrated their utility in classifying stress and non-stress conditions (Rashid et al. 2023). This study builds upon these foundations by integrating multiple machines learning models and enhancing their performance through deep learning approaches, making the system more robust and scalable
  • 6.
    06/13/2025 TABLE 1 -COMPARISON WITH EXISTING WORK Reference Methods Objective Rashid et al. 2022 Wearable Devices Stress Detection using Context Aware Sensor Fusion. Ninh et al. 2022 Wearable Devices Independent Stress Detection Model. Xu et al. 2024 SELF-CARE Detect Stress Using Physiological Signals Rashid et al. 2022 SELF-CARE Aware Low-Power Edge Computing for Stress Detection Vigouroux et al. 2025 Machine Learning Heart RateVariability-Based Mental Stress Detection Kuttala et al. 2023 Multimodal CNN Feature Fusion for Stress Detection 6
  • 7.
    06/13/2025 METHODOLOGY In this projectwe are implementing both software and hardware sides so we can increase the efficiency of our system we are going to collect important data like heartbeat, SPO2, GSR with the help of hardware like wearables and going to put the values in our software to track mental health which is going to increase overall efficiency. 1. Data Collection Wearable sensors capable of tracking the following physiological data were used to collect data: Movement patterns: 3-axis acceleration [ACC]. Electrocardiogram [ECG] to track heart activity. Blood Volume Pulse [BVP] to monitor changes in blood flow. BODY TEMP [TEMP] to capture changes in stress responses. Respiration [RESP] to quantify variations in the rate of breathing. Electromyogram [EMG] to monitor muscle activity. Electrodermal Activity [EDA], which when it comes to sweat gland activation, how much conductivity we get out of skin and shown in Fig.1. 7
  • 8.
    06/13/2025 2. Feature Extraction Severalfeatures contributing to detect mental health stress are physiological features, Behavioural features, contextual features they help us to detect many important parameters Like Respiration Rate, Heartrate, Electrodermal activity and other parameters like sleep pattern, speech features facial expression and less evolving parameters like social interaction and environmental factors (Ali et al. 2021). Preprocessing was performed on the collected data for noise removal and value normalization. Input vectors were constructed by extracting statistical features such as mean, variance, and peak-to-peak differences for each imaging channel in each time series interval (Öztekin et al. 2025). 3. Machine Learning Models The algorithms used for classification consisted of: K-Nearest Neighbour [KNN]: Useful for non-linear data classification linear discriminant analysis [LDA]: For dimensional reduction and class separation. Decision tree [DT]: For 0/1 classification, easy to understand. AdaBoost [AB] — Boosting weak classifiers. y=wTx (1) Note - That the equation is used for linear discriminant analysis it is use to simplify the high dimensional data into lower dimensional data. 8
  • 9.
    06/13/2025 4. Deep LearningApproach We employed Artificial Neural Networks [ANN] to model complex interdependencies in the data and obtained almost perfect classification of stress and non-stress conditions (Srinivasan et al. 2023). Cognitive Symptoms: Constant Depression, Harmful thoughts, ADHD Symptoms, OCD Symptoms, Inability to focus, Financial Problem, Being pessimistic. Physical Symptoms: Constipation or diarrhea, frequent colds and flu, aches and pains, nausea, and dizziness discomfort in the chest, Fast heartbeat, loose intestine, choking sensation, teeth grinding, frequent and urgent urination, fatigue, and weight loss or gain (Srivastava et al. 2019). Emotional Symptoms: Tension, irritability or anger, restlessness, worries, difficulty relaxing, depression, general dissatisfaction, anxiety and agitation, and moodiness are examples of emotional symptoms (Trivedi et al. 2022). Figure 1. - Process execution Conductance [Source: Researchgate.net] In figure 1. - Discussed machine learning models which includes BVP Which Measures Blood Volume Pulse, EDA Use for Measure Skin Conductance as Well as Skin temperature And Skin. 9
  • 10.
    06/13/2025 5. SYSTEM DESIGN Aswe already discuss that this project work on both aspects' software and hardware: Dashboard It is the user interface provides every information about the website the things it contains. Log In Allows users to create an account within the system. Visualization Provides visual representations of the user's mental health data, allowing them to track their metrics over time. (Shanmugasundaram et al. 2019). Wearables This includes smart watches, health bands, medical equipment Like Blood pressure machine etc. 10
  • 11.
    06/13/2025 Physiology Sensors ECG, GSR,PPG, SPO2, EMG And Temperature Sensor. • Behavioral Sensors Accelometer, Gyroscope, Camera Module, Arduino. Microphone, Wi-Fi Module, Battery pack etc. Figure 2. - No. of student vs. Type of Sensor In figure 2. - as you can see that it shows the number of students involved in to track their mental health using wearables which can record all the important vitals which are shown above It also shows the unavailability of these services in many areas. 11 EDA ECG EEG PPG EMG fNIRS RESP SKT MOVE PUPIL SpO2 BP 0 5 10 15 20 Multimodal Stuides Unimodal Stuides Total Studies Type of sensor Number of Student
  • 12.
    06/13/2025 Figure 3. –System Flow Chart 12 In figure 3. - Discussed about important vitals that are needed to track mental health. which includes several sensors like GSR, SPO2, ECG with the help of machine learning using decision tree.
  • 13.
    06/13/2025 RESULTS AND DISCUSSION Asshown in Fig. 4, this project works on hardware and software aspect. In hardware aspect we use sensors like - Fitbit, Apple Watch, GSS Sensors, Face detection to monitor user data which is going to help us to determine the Stress level where as in software aspect we don’t use or own any hardware we just input all the important data in the software and calculate the stress level of user but the problem is accuracy in software aspect we have to comprise with accuracy but it is an cheap alternative so for now we mainly depend on the software aspect of this project and in future we can also integrate hardware to increase the accuracy. As for now, in this era of new technologies people are concerned about their mental health, but they still don’t want to spend money on this type of hardware equipment so let’s continue with the software Aspect. These results punctuate the trustability of integrating wearable detectors with machine literacy algorithms for real- time stress monitoring. Compared to former styles that reckoned solely on single physiological signals. This System Demonstrates Bettered Robustness by integrating Multiple Futures Contemporaneously. These results punctuate the trustability of integrating wearable detectors with machine literacy algorithms for real- time stress monitoring. 13
  • 14.
    06/13/2025 14 Figure 4.- User Interface In figure 4. - In this page the User Supposed to Fill his/her important vital which he/her record with the help of wearables sensors. After entering all the vitals accurately. It shows you the result whereas this page records all the previous data of the user so the user can record his data which help him/her in overall well being.
  • 15.
    06/13/2025 CONCLUSION AND FUTUREWORK The Our system collects real-time physiological data from the user using various sensors that measure: Heart Rate & Blood Pressure – Indicators of stress, anxiety, or calmness. This data is then processed by our software, which analyses patterns and provides insights into the person’s mental state. Currently, our system achieves an accuracy rate of around 80-95%, and we are working on making it even more precise. This Paper presents a scalable and accurate stress detection system leveraging wearable sensor data and Integrated machine learning to formulate new ideas to control stress. The system surpasses existing models in accuracy and flexibility, making it suitable for real-world applications such as workplace stress management and health monitoring. Future research can focus on integrating facial recognition and speech analysis to complement physiological data, enhancing the accuracy further. Additionally, deploying this system in real-time smart phones software can enable broader adoption and continuous monitoring system in real-time mobile applications can enable broader adoption and continuous monitoringenhancing the accuracy further. Additionally, deploying this system in real-time mobile software can enable broader adoption and continuous monitoring. Mental health tracking should not be limited to clinical settings or expensive devices. Our goal is to make this technology affordable and accessible so that more people can monitor and manage their mental health proactively. By bridging the gap between biometric data and mental well-being, we hope to contribute to a world where mental health care is more effective, personalized, and within everyone’s reach. 15
  • 16.
    06/13/2025 REFERENCE 1. Banerjee, J.S.,Mahmud, M., & Brown, D. 2023. Heart rate variability-based mental stress detection: an explainable machine learning approach. SN Computer Science 4(2): 176. 2. Gedam, S., & Paul, S. 2021. A review on mental stress detection using wearable sensors and machine learning techniques. IEEE Access 9: 84045-84066. 3. Kuttala, R., Subramanian, R., & Oruganti, V.R.M. 2023. Multimodal hierarchical CNN feature fusion for stress detection. IEEE Access 11: 6867-6878. 4. Le Vigouroux, S., Chevrier, B., Montalescot, L., & Charbonnier, E. 2025. Post-pandemic student mental health and coping strategies: A time trajectory study. Journal of Affective Disorders. 5. M.S.F. Ali, M.A. Khan, & M.H. Rehmani. 2021. Wearable Sensor-Based Stress Detection Using Deep Learning: A Survey. IEEE Access 9: 12423-12441. 16
  • 17.
    06/13/2025 REFERENCE 6. Ninh, V.T.,Nguyen, M.D., Smyth, S., Tran, M.T., Healy, G., Nguyen, B.T., & Gurrin, C. 2022. An improved subject-independent stress detection model applied to consumer-grade wearable devices. In International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, pp. 907-919. Cham: Springer International Publishing. 7. Öztekin, G.G., Gómez-Salgado, J., & Yıldırım, M. 2025. Future anxiety, depression and stress among undergraduate students: psychological flexibility and emotion regulation as mediators. Frontiers in Psychology 16: 1517441. 8. Rashid, N., Mortlock, T., & Al Faruque, M.A. 2022. Self-care: Selective fusion with context-aware low-power edge computing for stress detection. In 2022 18th International Conference on Distributed Computing in Sensor Systems (DCOSS), pp. 49-52. IEEE. 9. Rashid, N., Mortlock, T., & Al Faruque, M.A. 2023. Stress detection using context-aware sensor fusion from wearable devices. IEEE Internet of Things Journal 10(16): 14114-14127. 10. Shanmugasundaram, G., Yazhini, S., Hemapratha, E., & Nithya, S. 2019. A comprehensive review on stress detection techniques. In 2019 IEEE International Conference on System, Computation, Automation and Networking (ICSCAN), pp. 1-6. IEEE. 17
  • 18.
    06/13/2025 REFERENCE 11. Trivedi, N.K.,Tiwari, R.G., Witarsyah, D., Gautam, V., Misra, A., & Nugraha, R.A. 2022. Machine learning based evaluations of stress, depression, and anxiety. In 2022 International Conference Advancement in Data Science, E-learning and Information Systems (ICADEIS), pp. 1-5. IEEE. 12. Xu, G., Qin, R., Zheng, Z., & Shi, Y. 2024. An Adaptive System for Wearable Devices to Detect Stress Using Physiological Signals. arXiv preprint arXiv:2407.15252. 18
  • 19.
  • 20.