Notícias

Banca de QUALIFICAÇÃO: FRANCISCO ÉRBIO DIAS

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
DISCENTE: FRANCISCO ÉRBIO DIAS
DATA: 18/08/2023
HORA: 09:00
LOCAL: Bloco 8 - Centro de Tecnologia
TÍTULO: Classification of vehicle data in a telemetry system using Machine Learning with AutoML
PALAVRAS-CHAVES: Telemetry, AutoML, Hyperparameter, Machine Learning
PÁGINAS: 15
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Elétrica
SUBÁREA: Eletrônica Industrial, Sistemas e Controles Eletrônicos
ESPECIALIDADE: Controle de Processos Eletrônicos, Retroalimentação
RESUMO:

An intelligent telemetry system is a valuable tool for collecting and monitoring real-time data from a variety of devices and sensors, where optimizing the parameters of the data classification model is an important process for improving system performance. Instead of manually defining these parameters, optimization seeks to automatically find the ideal values by minimizing a performance metric. However, in a telemetry system, the large amount of generated data can make manual analysis difficult, requiring automated solutions. In this context, the use of machine learning techniques with Automated Machine Learning (AutoML) has proven to be an efficient option for optimizing models and analyzing a large volume of data in an automated manner. This study utilizes four machine learning techniques: K-NN, Multi-Layer Perceptron (MLP) Artificial Neural Networks, Random Forest algorithm, and Extreme Gradient Boosting with crossvalidation and the k-fold technique, with 10 folds applied to two vehicular databases obtained from a telemetry system. The results of the hyperparameter optimization algorithms showed superior performance compared to results without the use of AutoML techniques, in terms of accuracy rates across all applied machine learning techniques.


MEMBROS DA BANCA:
Interno - 1785470 - JOSE MARIA PIRES DE MENEZES JUNIOR
Interno - 1644191 - JOSE MEDEIROS DE ARAUJO JUNIOR
Interno - 289819 - OTACILIO DA MOTA ALMEIDA
Notícia cadastrada em: 19/08/2023 09:16
SIGAA | Superintendência de Tecnologia da Informação - STI/UFPI - (86) 3215-1124 | © UFRN | sigjb05.ufpi.br.instancia1 23/07/2024 20:51