Modelos de predição aplicados ao aprendizado motor

• Ano de defesa: 2019
• Autor(a) principal: SANTOS, Moisés Rocha dos
• Orientador(a): ALMEIDA NETO, Areolino de
• Banca de defesa: ALMEIDA NETO, Areolino de , RIBEIRO, Paulo Rogério de Almeida , BARRADAS FILHO, Alex Oliveira , BRASIL, Fabrício Lima
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Maranhão
• Programa de Pós-Graduação: PROGRAMA DE PÓS-GRADUAÇÃO EM CIÊNCIA DA COMPUTAÇÃO/CCET
• Departamento: DEPARTAMENTO DE INFORMÁTICA/CCET
• País: Brasil
• Palavras-chave em Português: Tarefa de traçado; Aquisição de habilidade motora; Modelos de classificação
• Palavras-chave em Inglês: Tracing ask; Motor skill learning; Classification models
• Área do conhecimento CNPq: Teoria da Computação; Análise de Algoritmos e Complexidade de Computação; Processos Perceptuais e Motores
• Link de acesso: https://tedebc.ufma.br/jspui/handle/tede/2571

The present work aims to propose an approach to estimate the number of sessions required to learn a motor task. Motor activities are the main way of interacting with the world. Therefore, loss of ability to perform some of these activities as a result of a neurological disease is a serious injury to the individual. In the literature, there are many works on motor learning, mostly looking for ways to decrease the time of skill acquisition or motor rehabilitation. However, few works concentrate on trying to estimate the training time needed to achieve certain motor performance. The methodology consisted of a review of the state of the art of motor skill acquisition, as well as the initial configuration of a training platform, the application of a pilot experiment with three participants and a final experiment with eight participants. In the pilot experiment, a three-block training session for each participant was performed and it aimed to predict in which block the participant was. From three real participants, 18 simulated participants were generated, in order to measure the performance of the experiment with more participants, and the block was estimated through the average performance of the participants. In the final experiment, three sessions were performed for each participant, whose purpose was to predict in which session the participant would reach a certain error based on their profile and initial performance. The classification models used in the final experiment were: Algorithm K-Neighbors Nearer, Neural Network, Decision Tree, Support Vector Machine and Automatic Machine Learning (AutoML) with "Auto Weka". In the results of the pilot experiment, an improvement of motor skills was observed after the training. Through the data from the pilot experiment, the best results were obtained using the Decision Tree algorithm. In the results of the final experiment, it was possible to observe the motor improvement and the consistency. Using the data from the final experiment, the best results were obtained with AutoML. The work showed the possibility of estimating the number of sessions to achieve a certain performance using prediction algorithms. In addition, the relevance of the work is accentuated, since this will serve as a basis for future experiments with more healthy participants, as well as people with motor damage.