This article aims to classify texts and predict the categories of occurrences, through the study of Artificial Intelligence models, using Machine Learning and Deep Learning for the classification of texts and analysis of predictions, suggesting the best option with the smallest error.
The solution was designed to be implemented in two stages: Machine Learning and Application, according to the diagram below from the Data Science Academy.
Hot Topics in Machine Learning for Research and ThesisWriteMyThesis
Machine Learning is a hot topic for research for research. There are various good thesis topics in Machine Learning. Writemythesis provides thesis in Machine Learning along with proper guidance in this field. Find the list of thesis topics in this document.
http://www.writemythesis.org/master-thesis-topics-in-machine-learning/
Deep learning algorithms have drawn the attention of researchers working in the field of computer vision, speech
recognition, malware detection, pattern recognition and natural language processing. In this paper, we present an overview of
deep learning techniques like Convolutional neural network, deep belief network, Autoencoder, Restricted Boltzmann machine
and recurrent neural network. With this, current work of deep learning algorithms on malware detection is shown with the
help of literature survey. Suggestions for future research are given with full justification. We also showed the experimental
analysis in order to show the importance of deep learning techniques.
Hot Topics in Machine Learning for Research and ThesisWriteMyThesis
Machine Learning is a hot topic for research for research. There are various good thesis topics in Machine Learning. Writemythesis provides thesis in Machine Learning along with proper guidance in this field. Find the list of thesis topics in this document.
http://www.writemythesis.org/master-thesis-topics-in-machine-learning/
Deep learning algorithms have drawn the attention of researchers working in the field of computer vision, speech
recognition, malware detection, pattern recognition and natural language processing. In this paper, we present an overview of
deep learning techniques like Convolutional neural network, deep belief network, Autoencoder, Restricted Boltzmann machine
and recurrent neural network. With this, current work of deep learning algorithms on malware detection is shown with the
help of literature survey. Suggestions for future research are given with full justification. We also showed the experimental
analysis in order to show the importance of deep learning techniques.
Gives an Introduction to Deep learning, What can you achieve with deep learning. What is deep learning's relationship with machine learning. Technical basics of working of deep learning. Introduction to LSTM. How LSTM can be used for Text classification. Results obtained.. Practical recommendations.
FACE EXPRESSION RECOGNITION USING CONVOLUTION NEURAL NETWORK (CNN) MODELS ijgca
This paper proposes the design of a Facial Expression Recognition (FER) system based on deep
convolutional neural network by using three model. In this work, a simple solution for facial expression
recognition that uses a combination of algorithms for face detection, feature extraction and classification
is discussed. The proposed method uses CNN models with SVM classifier and evaluates them, these models
are Alex-net model, VGG-16 model and Res-Net model. Experiments are carried out on the Extended
Cohn-Kanada (CK+) datasets to determine the recognition accuracy for the proposed FER system. In this
study the accuracy of AlexNet model compared with Vgg16 model and ResNet model. The result show that
AlexNet model achieved the best accuracy (88.2%) compared to other models.
Traditional Machine Learning had used handwritten features and modality-specific machine learning to classify images, text or recognize voices. Deep learning / Neural network identifies features and finds different patterns automatically. Time to build these complex tasks has been drastically reduced and accuracy has exponentially increased because of advancements in Deep learning. Neural networks have been partly inspired from how 86 billion neurons work in a human and become more of a mathematical and a computer problem. We will see by the end of the blog how neural networks can be intuitively understood and implemented as a set of matrix multiplications, cost function, and optimization algorithms.
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHESVikash Kumar
IMAGE CLASSIFICATION USING KNN, RANDOM FOREST AND SVM ALGORITHM ON GLAUCOMA DATASETS AND EXPLAIN THE ACCURACY, SENSITIVITY, AND SPECIFICITY OF EACH AND EVERY ALGORITHMS
Mengenal Machine/Deep Learning, Artificial Intelligence dan mengenal apa bedanya dengan Business Intelligence, apa hubungannya dengan Big Data dan Data Science/Analytics.
O presente artigo tem como objetivo o estudo de modelos de Inteligência Artificial para sistemas de recomendação, case Santander do Kaggle, utilizando algoritmos de Machine Learning e Deep Learning para fazer a recomendação de produtos com base na história dos clientes do banco.
This article aims to contribute to a population service strategy in the fight against Covid-19, through a scenario simulator with Artificial Intelligence and System Dynamics.
In February 2021, relevant changes were announced in Brazil in the fight against Covid-19, especially regarding vaccination. This article addresses a series of predictions to assist in the strategy of serving the population in the fight against this disease. It is a good opportunity to better understand the "invisible" factors that affect the people's contamination.
To help model these "invisible" factors, social distancing data was obtained, this data was passed as input to an Artificial Intelligence system and predictions were performed in Deep Learning (LSTM). The prediction of the social distancing variable was exported to calibrate the model in System Dynamics, with the Stella software, and the simulations were made with a Covid-19 model.
Gives an Introduction to Deep learning, What can you achieve with deep learning. What is deep learning's relationship with machine learning. Technical basics of working of deep learning. Introduction to LSTM. How LSTM can be used for Text classification. Results obtained.. Practical recommendations.
FACE EXPRESSION RECOGNITION USING CONVOLUTION NEURAL NETWORK (CNN) MODELS ijgca
This paper proposes the design of a Facial Expression Recognition (FER) system based on deep
convolutional neural network by using three model. In this work, a simple solution for facial expression
recognition that uses a combination of algorithms for face detection, feature extraction and classification
is discussed. The proposed method uses CNN models with SVM classifier and evaluates them, these models
are Alex-net model, VGG-16 model and Res-Net model. Experiments are carried out on the Extended
Cohn-Kanada (CK+) datasets to determine the recognition accuracy for the proposed FER system. In this
study the accuracy of AlexNet model compared with Vgg16 model and ResNet model. The result show that
AlexNet model achieved the best accuracy (88.2%) compared to other models.
Traditional Machine Learning had used handwritten features and modality-specific machine learning to classify images, text or recognize voices. Deep learning / Neural network identifies features and finds different patterns automatically. Time to build these complex tasks has been drastically reduced and accuracy has exponentially increased because of advancements in Deep learning. Neural networks have been partly inspired from how 86 billion neurons work in a human and become more of a mathematical and a computer problem. We will see by the end of the blog how neural networks can be intuitively understood and implemented as a set of matrix multiplications, cost function, and optimization algorithms.
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHESVikash Kumar
IMAGE CLASSIFICATION USING KNN, RANDOM FOREST AND SVM ALGORITHM ON GLAUCOMA DATASETS AND EXPLAIN THE ACCURACY, SENSITIVITY, AND SPECIFICITY OF EACH AND EVERY ALGORITHMS
Mengenal Machine/Deep Learning, Artificial Intelligence dan mengenal apa bedanya dengan Business Intelligence, apa hubungannya dengan Big Data dan Data Science/Analytics.
O presente artigo tem como objetivo o estudo de modelos de Inteligência Artificial para sistemas de recomendação, case Santander do Kaggle, utilizando algoritmos de Machine Learning e Deep Learning para fazer a recomendação de produtos com base na história dos clientes do banco.
This article aims to contribute to a population service strategy in the fight against Covid-19, through a scenario simulator with Artificial Intelligence and System Dynamics.
In February 2021, relevant changes were announced in Brazil in the fight against Covid-19, especially regarding vaccination. This article addresses a series of predictions to assist in the strategy of serving the population in the fight against this disease. It is a good opportunity to better understand the "invisible" factors that affect the people's contamination.
To help model these "invisible" factors, social distancing data was obtained, this data was passed as input to an Artificial Intelligence system and predictions were performed in Deep Learning (LSTM). The prediction of the social distancing variable was exported to calibrate the model in System Dynamics, with the Stella software, and the simulations were made with a Covid-19 model.
O presente artigo tem como objetivo a classificação de textos e a previsão das categorias de ocorrências, através do estudo de modelos de Inteligência Artificial, utilizando Machine Learning e Deep Learning para a classificação de textos e análise das previsões, sugerindo-se a melhor opção com o menor erro.
O presente artigo tem como objetivo contribuir para uma estratégia de atendimento à população no combate à Covid-19, através de um simulador de cenários com Inteligência Artificial e Dinâmica de Sistemas.
Em fevereiro de 2021 foram anunciadas no Brasil mudanças relevantes no combate à Covid-19, especialmente quanto à vacinação. Aborda-se nesse artigo uma série de previsões para auxiliar na estratégia de atendimento à população no combate a esta doença, é uma boa oportunidade para entender melhor os fatores "invisíveis" que afetam a contaminação das pessoas.
Para ajudar a modelar esses fatores "invisíveis", partiu-se da obtenção de dados de distanciamento social, esses dados foram passados como entrada em um sistema de Inteligência Artificial e foram executadas previsões em Deep Learning (LSTM). A previsão da variável distanciamento social foi exportada para calibrar o modelo em Systems Dynamics, com o software Stella, e as simulações foram feitas com um modelo de Covid-19. Como trabalho futuro, o resultado dessas simulações serão plotados em um mapa de um telefone celular através de geoprocessamento.
We are living in an era where changes are taking place very quickly.
Imagine a scenario where we can use Machine Learning to identify changes that are associated with incidents and problems.
Then classify them using the Cynefin framework to determine the context of change according to the 5 domains: simple, complicated, complex, chaotic and disorder (natural science).
Come up with a Decision Intelligence model that can provide the best answer to the potential problems caused by this ever evolving environment.
Finally add the Requests to the Incident Model, as ITSM focus, on IT4IT Value Stream Network!
Using Machine Learning embedded in Organizational Responsibility Model, added to the ten characteristics of the CIO Master and the twelve competencies of the workforce can help lead the Digital Transformation of the traditional public organizations to the Exponential.
We are living in an era where changes are taking place very quickly.
Imagine a scenario where we can use Machine Learning to identify changes that are associated with problems.
Then classify them using the Cynefin framework to determine the context of change according to the 5 domains: simple, complicated, complex, chaotic and disorder (natural science).
And finally come up with a Decision Intelligence model that can provide the best answer to the potential problems caused by this ever evolving environment!
O uso de Machine Learning incorporado no Modelo de Responsabilidade Organizacional, agregado às dez características do Mestre CIO e às doze competências da força de trabalho pode ajudar a liderar as Transformações Digitais das organizações públicas do tradicional para o Exponencial.
Objetivando preparar as líderes e organizações para o futuro (do tradicional para o exponencial), o Modelo de Responsabilidade Organizacional foi adaptado para abranger além do contexto analítico, com Inteligência Artificial e Aprendizagem de Máquina, o contexto mais subjetivo com Dez Características do Mestre CIO (por Álvaro Mello), bem como as doze Competências da Força de Trabalho (por Lily Mok).
Como o Aprendizado de Máquina pode nos ajudar a liderar as transformações das organizações públicas do tradicional para o Exponencial? (parte 2)
Superamos o desafio de substituir os dados dessas 1.000 simulações por dados reais de uma empresa de TI e inseri-los na Árvore de Decisão. Este modelo tem a capacidade de prever qual é o melhor Modal (TI Bimodal) para realizar determinada atividade: Tradicional ou Exponencial.
Estamos usando o Aprendizado de Máquina desenvolvido no ambiente Notebook Jupyter, na linguagem Python, para auxiliar na decisão de quando usar o tradicional ou exponencial.
Como o Aprendizado de Máquina pode nos ajudar a liderar as transformações das organizações públicas do tradicional para o Exponencial? (parte 1)
Primeiro adaptamos o Modelo de Responsabilidade Organizacional para ser Exponencial com o NetLogo, depois com alguns algoritmos de Aprendizado de Máquina nós construímos um Modelo de Árvore de Decisão que tem como entrada dados de 1000 simulações.
Estamos usando o Aprendizado de Máquina desenvolvido no ambiente Notebook Jupyter, na linguagem Python, para auxiliar na decisão de quando usar o tradicional ou exponencial.
Aplicação da dinâmica de Análise de Cenário, apresentada na Oficina ABEP-FACIN e utilizada pela PRODEMGE, para implantar uma solução corporativa de geoprocessamento no Estado de Minas Gerais, utilizando o FACIN.
O modelo mapeia o alinhamento entre o roteiro de recomendações, proposto pelo Decreto nº 47.154 que regulamenta a chamada de Lei das Estatais nº 13.303, e as boas práticas de governança corporativa elaboradas por Pedro Bramont e João Souza Neto para que possam ser aplicadas nas empresas públicas.
A Inteligência Artificial contribui para criar, ampliar e melhorar os sistemas de serviços, estabelecendo os algoritmos que irão subsidiar o processo de tomada de decisão no ambiente das organizações. Também contribuem para a solução dos problemas relacionados à gestão por resultados, visando à melhoria dos serviços públicos prestados, em face de uma demanda de serviços superior à capacidade instalada.
Alinhamento da Visão Negócios do FACIN e o Modelo de Responsabilidade Organizacional, iniciativas voltadas para a melhoria da gestão e governança nas organizações públicas.
Key Trends Shaping the Future of Infrastructure.pdfCheryl Hung
Keynote at DIGIT West Expo, Glasgow on 29 May 2024.
Cheryl Hung, ochery.com
Sr Director, Infrastructure Ecosystem, Arm.
The key trends across hardware, cloud and open-source; exploring how these areas are likely to mature and develop over the short and long-term, and then considering how organisations can position themselves to adapt and thrive.
Let's dive deeper into the world of ODC! Ricardo Alves (OutSystems) will join us to tell all about the new Data Fabric. After that, Sezen de Bruijn (OutSystems) will get into the details on how to best design a sturdy architecture within ODC.
State of ICS and IoT Cyber Threat Landscape Report 2024 previewPrayukth K V
The IoT and OT threat landscape report has been prepared by the Threat Research Team at Sectrio using data from Sectrio, cyber threat intelligence farming facilities spread across over 85 cities around the world. In addition, Sectrio also runs AI-based advanced threat and payload engagement facilities that serve as sinks to attract and engage sophisticated threat actors, and newer malware including new variants and latent threats that are at an earlier stage of development.
The latest edition of the OT/ICS and IoT security Threat Landscape Report 2024 also covers:
State of global ICS asset and network exposure
Sectoral targets and attacks as well as the cost of ransom
Global APT activity, AI usage, actor and tactic profiles, and implications
Rise in volumes of AI-powered cyberattacks
Major cyber events in 2024
Malware and malicious payload trends
Cyberattack types and targets
Vulnerability exploit attempts on CVEs
Attacks on counties – USA
Expansion of bot farms – how, where, and why
In-depth analysis of the cyber threat landscape across North America, South America, Europe, APAC, and the Middle East
Why are attacks on smart factories rising?
Cyber risk predictions
Axis of attacks – Europe
Systemic attacks in the Middle East
Download the full report from here:
https://sectrio.com/resources/ot-threat-landscape-reports/sectrio-releases-ot-ics-and-iot-security-threat-landscape-report-2024/
JMeter webinar - integration with InfluxDB and GrafanaRTTS
Watch this recorded webinar about real-time monitoring of application performance. See how to integrate Apache JMeter, the open-source leader in performance testing, with InfluxDB, the open-source time-series database, and Grafana, the open-source analytics and visualization application.
In this webinar, we will review the benefits of leveraging InfluxDB and Grafana when executing load tests and demonstrate how these tools are used to visualize performance metrics.
Length: 30 minutes
Session Overview
-------------------------------------------
During this webinar, we will cover the following topics while demonstrating the integrations of JMeter, InfluxDB and Grafana:
- What out-of-the-box solutions are available for real-time monitoring JMeter tests?
- What are the benefits of integrating InfluxDB and Grafana into the load testing stack?
- Which features are provided by Grafana?
- Demonstration of InfluxDB and Grafana using a practice web application
To view the webinar recording, go to:
https://www.rttsweb.com/jmeter-integration-webinar
"Impact of front-end architecture on development cost", Viktor TurskyiFwdays
I have heard many times that architecture is not important for the front-end. Also, many times I have seen how developers implement features on the front-end just following the standard rules for a framework and think that this is enough to successfully launch the project, and then the project fails. How to prevent this and what approach to choose? I have launched dozens of complex projects and during the talk we will analyze which approaches have worked for me and which have not.
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualityInflectra
In this insightful webinar, Inflectra explores how artificial intelligence (AI) is transforming software development and testing. Discover how AI-powered tools are revolutionizing every stage of the software development lifecycle (SDLC), from design and prototyping to testing, deployment, and monitoring.
Learn about:
• The Future of Testing: How AI is shifting testing towards verification, analysis, and higher-level skills, while reducing repetitive tasks.
• Test Automation: How AI-powered test case generation, optimization, and self-healing tests are making testing more efficient and effective.
• Visual Testing: Explore the emerging capabilities of AI in visual testing and how it's set to revolutionize UI verification.
• Inflectra's AI Solutions: See demonstrations of Inflectra's cutting-edge AI tools like the ChatGPT plugin and Azure Open AI platform, designed to streamline your testing process.
Whether you're a developer, tester, or QA professional, this webinar will give you valuable insights into how AI is shaping the future of software delivery.
Neuro-symbolic is not enough, we need neuro-*semantic*Frank van Harmelen
Neuro-symbolic (NeSy) AI is on the rise. However, simply machine learning on just any symbolic structure is not sufficient to really harvest the gains of NeSy. These will only be gained when the symbolic structures have an actual semantics. I give an operational definition of semantics as “predictable inference”.
All of this illustrated with link prediction over knowledge graphs, but the argument is general.
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf91mobiles
91mobiles recently conducted a Smart TV Buyer Insights Survey in which we asked over 3,000 respondents about the TV they own, aspects they look at on a new TV, and their TV buying preferences.
The Art of the Pitch: WordPress Relationships and SalesLaura Byrne
Clients don’t know what they don’t know. What web solutions are right for them? How does WordPress come into the picture? How do you make sure you understand scope and timeline? What do you do if sometime changes?
All these questions and more will be explored as we talk about matching clients’ needs with what your agency offers without pulling teeth or pulling your hair out. Practical tips, and strategies for successful relationship building that leads to closing the deal.
LF Energy Webinar: Electrical Grid Modelling and Simulation Through PowSyBl -...DanBrown980551
Do you want to learn how to model and simulate an electrical network from scratch in under an hour?
Then welcome to this PowSyBl workshop, hosted by Rte, the French Transmission System Operator (TSO)!
During the webinar, you will discover the PowSyBl ecosystem as well as handle and study an electrical network through an interactive Python notebook.
PowSyBl is an open source project hosted by LF Energy, which offers a comprehensive set of features for electrical grid modelling and simulation. Among other advanced features, PowSyBl provides:
- A fully editable and extendable library for grid component modelling;
- Visualization tools to display your network;
- Grid simulation tools, such as power flows, security analyses (with or without remedial actions) and sensitivity analyses;
The framework is mostly written in Java, with a Python binding so that Python developers can access PowSyBl functionalities as well.
What you will learn during the webinar:
- For beginners: discover PowSyBl's functionalities through a quick general presentation and the notebook, without needing any expert coding skills;
- For advanced developers: master the skills to efficiently apply PowSyBl functionalities to your real-world scenarios.
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024Tobias Schneck
As AI technology is pushing into IT I was wondering myself, as an “infrastructure container kubernetes guy”, how get this fancy AI technology get managed from an infrastructure operational view? Is it possible to apply our lovely cloud native principals as well? What benefit’s both technologies could bring to each other?
Let me take this questions and provide you a short journey through existing deployment models and use cases for AI software. On practical examples, we discuss what cloud/on-premise strategy we may need for applying it to our own infrastructure to get it to work from an enterprise perspective. I want to give an overview about infrastructure requirements and technologies, what could be beneficial or limiting your AI use cases in an enterprise environment. An interactive Demo will give you some insides, what approaches I got already working for real.
PHP Frameworks: I want to break free (IPC Berlin 2024)Ralf Eggert
In this presentation, we examine the challenges and limitations of relying too heavily on PHP frameworks in web development. We discuss the history of PHP and its frameworks to understand how this dependence has evolved. The focus will be on providing concrete tips and strategies to reduce reliance on these frameworks, based on real-world examples and practical considerations. The goal is to equip developers with the skills and knowledge to create more flexible and future-proof web applications. We'll explore the importance of maintaining autonomy in a rapidly changing tech landscape and how to make informed decisions in PHP development.
This talk is aimed at encouraging a more independent approach to using PHP frameworks, moving towards a more flexible and future-proof approach to PHP development.
Builder.ai Founder Sachin Dev Duggal's Strategic Approach to Create an Innova...Ramesh Iyer
In today's fast-changing business world, Companies that adapt and embrace new ideas often need help to keep up with the competition. However, fostering a culture of innovation takes much work. It takes vision, leadership and willingness to take risks in the right proportion. Sachin Dev Duggal, co-founder of Builder.ai, has perfected the art of this balance, creating a company culture where creativity and growth are nurtured at each stage.
FIDO Alliance Osaka Seminar: Passkeys at Amazon.pdf
Occurrence Prediction_NLP
1. Text classification with Deep Learning
Occurrence forecasting models
Guttenberg Ferreira Passos
This article aims to classify texts and predict the categories of occurrences, through the study
of Artificial Intelligence models, using Machine Learning and Deep Learning for the
classification of texts and analysis of predictions, suggesting the best option with the smallest
error.
The solution was designed to be implemented in two stages: Machine Learning and
Application, according to the diagram below from the Data Science Academy.
Source: Data Science Academy
The focus of this article is the Machine Learning stage, the application development is out of scope and
may be the object of future work.
The solution was applied to an agency in the State of Minas Gerais with the collection of three data sets
containing 5,000, 100,000 and 1,740,000 occurrences respectively.
The project of elaboration of the algorithms of the Machine Learning stage was divided into four phases:
1) Development of a prototype for customer approval, with the Orange tool, for training a sample
of 5,000 occurrences and forecasting 300 occurrences. In this step, Machine Learning
algorithms were used.
2) Development of a Python program for training a sample of 100,000 occurrences and
forecasting 300 occurrences. In this step, Deep Learning algorithms were used.
3) Training of a sample of 1,700,000 occurrences and prediction of 300 occurrences, using the
same environment.
4) Training of a sample of 1,700,000 occurrences and forecast of 60,000 occurrences, using the
same environment.
All models were adapted from the website https://orangedatamining.com/, of the videos:
https://www.youtube.com/watch?v=HXjnDIgGDuI&t=10s&ab_channel=OrangeDataMining and the Data
Science Academy Deep Learning II course classes: https://www.datascienceacademy.com.br
2. Machine Learning Algorithms Used:
AdaBoost
kNN
Logistic Regression
Naive Bayes
Random Forest
Deep Learning Algorithms Used:
LSTM - Long short-term memory
GRU - Gated Recurrent Unit
CNN - Convolutional Neural Networks
AdaBoost
It is a machine learning algorithm, derived from Adaptive Boosting. AdaBoost is adaptive in the sense
that subsequent ratings made are adjusted in favor of instances negatively rated by previous ratings.
AdaBoost is sensitive to noise in data and isolated cases. However, for some problems it is less
susceptible to loss of generalization ability after learning many training patterns (overfitting) than most
machine learning algorithms.
kNN
It is a machine learning algorithm, the kNN algorithm looks for the nearest k training examples in the
feature space and uses their mean as a prediction.
Logistic Regression
The logistic regression classification algorithm with LASSO regularization (L1) or crest (L2). Logistic
regression learns a logistic regression model from the data. It only works for sorting tasks.
Naive Bayes
A fast and simple probabilistic classifier based on Bayes' theorem with the assumption of feature
independence. It only works for sorting tasks.
Random Forest
Random Forest builds a set of decision trees. Each tree is developed from a bootstrap sample of training
data. When developing individual trees, an arbitrary subset of attributes is drawn (hence the term
“Random”), from which the best attribute for splitting is selected. The final model is based on the
majority of votes from individually grown trees in the forest..
Source of Machine Learning Algorithms: Wikipedia and https://orange3.readthedocs.io/en/latest
LSTM
The Long Short Term Memory - LSTM network is a recurrent neural network, which is used in several
Natural Language Processing scenarios. LSTM is a recurrent neural network (RNN) architecture that
“remembers” values at arbitrary intervals. LSTM is well suited for classifying, processing, and predicting
time series with time intervals of unknown duration. The relative gap length insensitivity gives LSTM an
advantage over traditional RNNs (also called “vanilla”), Hidden Markov Models (MOM) and other
sequence learning methods.
3. GRU
The Gated Recurrent Unit - GRU network aims to solve the problem of gradient dissipation that is
common in a standard recurrent neural network. The GRU can also be considered a variation of the
LSTM because both are similarly designed and in some cases produce equally excellent results.
CNN
The Convolutional Neural Network - CNN is a Deep Learning algorithm that can capture an input image,
assign importance (learned weights and biases) to various aspects/objects of the image and be able to
differentiate one from the other. The pre-processing required in a CNN is much less compared to other
classification algorithms. While in primitive methods filters are made by hand, with enough training,
CNNs have the ability to learn these filters/features..
Source of Deep Learning Algorithms: https://www.deeplearningbook.com.br
Neural networks are computing systems with interconnected nodes that function like neurons in the
human brain. Using algorithms, they can recognize hidden patterns and correlations in raw data, group
and classify them, and over time continually learn and improve.
The Asimov Institute https://www.asimovinstitute.org/neural-network-zoo/ published a cheat sheet
containing various neural network architectures, we will focus on the architectures highlighted in red
LSTM, GRU and CNN.
4. Source: THE ASIMOV INSTITUTE
Deep Learning is one of the foundations of Artificial Intelligence (AI), a type of machine learning
(Machine Learning) that trains computers to perform tasks like humans, which includes speech
recognition, image identification and predictions, learning over time. We can say that it is a Neural
Network with several hidden layers:
5. Phase 1
Phase 1 of the project is the development of a prototype for the presentation of the solution and its first
approval by the customer. The tool that was chosen for this phase is Orange Canvas, because it is a
more user-friendly graphical environment. In this environment, elements are dragged to the canvas
without having to type lines of code.
The work begins with Exploratory Data Analysis. Initially, it was found that the first sample of 5,000
occurrences was unbalanced, as shown below. It was decided to discard the occurrences of the
categories with the lowest volume of data.
The first phase was structured in three steps: pre-processing and data analysis, training the models and
forecasting the categories of occurrences. The stages were planned to facilitate the development and
implementation of the project as they are independent of each other and their processing is completed
at each stage, not needing to be repeated in the later stage.
Phase 1 - Step 1: Pre-processing and data analysis
In the first stage, data collection, pre-processing and analysis are carried out, as shown in the figure
below in the Orange tool.
Samples of 5,000 occurrences were collected to train the model and 300 occurrences to make the
prediction, simulating a production environment.
6. After collection, the data are organized into a Corpus to carry out the pre-processing, performing the
Transformation, Tokenization and Filtering actions of the data.
Words are arranged in a Bag of Words format, a simplified representation used in Natural Language
Processing - NLP. In this model, a text (such as a sentence or a document) is represented as the bag of its
words, disregarding grammar and even word order, but maintaining multiplicity.
Machine learning is a branch of artificial intelligence based on the idea that systems can learn from data,
identify patterns and make decisions with minimal human intervention. Machine learning explores the
study and construction of algorithms that can learn from their errors and make predictions about data.
Machine learning can be classified into two categories:
Supervised learning: The computer is presented with examples of desired inputs and outputs.
Unsupervised Learning: No tags are given to the learning algorithm, leaving it alone to find patterns in
the given inputs.
Through unsupervised learning it is possible to identify the Clusters and their hierarchy.
7. With multidimensional scaling (MDS) there is a way to visualize the level of similarity of individual cases
of a data set and the regions of the Clusters.
In addition, there is also an idea of the ease or difficulty of the model in making its predictions, the more
grouped the occurrences in a given region, the greater the probability of the model being correct.
Phase 1 - Step 2: Model training
The second step is to train the models using the following Machine Learning algorithms: AdaBoost, kNN,
Logistic Regression, Naive Bayes and Random Forest.
8. The overall performance of models can be measured through their Accuracy (AC), the proximity of a
result to its real reference value. Thus, the greater the accuracy, the closer to the reference or real value
is the result found.
The successes and errors identified in the result can be analyzed through the Confusion Matrix. On the
main diagonal of the matrix are the hits, correct predictions according to the real set. Errors are off the
main diagonal, incorrect predictions according to the real set.
Phase 1 - Step 3: Forecasting the categories of occurrences
The last stage of the prototype, phase 1 of the project, is the prediction of the categories of occurrences,
performed by each machine learning algorithm.
9. The result can be obtained through the probabilistic classification of observations, characterizing them
in pre-defined classes. The predicted class will be the one with the highest probability:
Phases 2, 3 and 4
For phases 2, 3 and 4 of the project, programs were developed in Python language for analysis, training
and prediction of occurrences, using the following Deep Learning algorithms: LSTM, GRU and CNN.
Samples of 100,000 and 1,700,000 occurrences were provided for training and for prediction samples of
300 and 60,000 occurrences, using the same environment.
The new hit samples were preprocessed and balanced:
10. The programs developed were structured respecting the same three steps as in the previous phase: pre-
processing and data analysis, training the models and forecasting the categories of occurrences.
In step 1, several data pre-processing techniques were used, similar to those used in the Orange Canvas
environment.
In the second stage, different architectures were developed for each Deep Learning algorithm.
Model 1 LSTMs - Neural Network Layers:
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding (Embedding) (None, 250, 100) 5000000
_________________________________________________________________
lstm (LSTM) (None, 100) 80400
_________________________________________________________________
dense (Dense) (None, 6) 606
=================================================================
Total params: 5,081,006
Trainable params: 5,081,006
Non-trainable params: 0
Model 2 LSTMs and CNNs - Neural Network Layers:
Model: "sequential_1"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding_1 (Embedding) (None, 250, 100) 5000000
_________________________________________________________________
conv1d (Conv1D) (None, 250, 32) 9632
_________________________________________________________________
max_pooling1d (MaxPooling1D) (None, 125, 32) 0
_________________________________________________________________
lstm_1 (LSTM) (None, 125, 100) 53200
12. Model 5 GRU and CNN - Neural Network Layers:
Model: "sequential_2"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding_2 (Embedding) (None, 250, 100) 5000000
_________________________________________________________________
conv1d (Conv1D) (None, 250, 32) 9632
_________________________________________________________________
max_pooling1d (MaxPooling1D) (None, 125, 32) 0
_________________________________________________________________
gru_1 (GRU) (None, 125, 100) 40200
_________________________________________________________________
gru_2 (GRU) (None, 100) 60600
_________________________________________________________________
dense_1 (Dense) (None, 6) 606
=================================================================
Total params: 5,111,038
Trainable params: 5,111,038
Non-trainable params: 0
Model 6 GRU with Dropout - Neural Network Layers:
Model: "sequential_3"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding_3 (Embedding) (None, 250, 100) 5000000
_________________________________________________________________
gru_3 (GRU) (None, 250, 200) 181200
_________________________________________________________________
gru_4 (GRU) (None, 200) 241200
_________________________________________________________________
dense_2 (Dense) (None, 6) 1206
=================================================================
Total params: 5,423,606
Trainable params: 5,423,606
Non-trainable params: 0
Conclusion
In this work, without any pretension of exhausting the subject, it was demonstrated that the models
based on Deep Learning had a better result than the other algorithms, as shown below:
13. The performance achieved by the combination of the LSTM and CNN algorithms is considered excellent,
with an accuracy of 97%. It is therefore recommended to adopt this model for the development of the
Occurrence Forecast application in production.