Evaluation of Training Executed by Web Using Fuzzy Rule Based Expert Systems

1 Ronei Marcos de Moraes, Department of Statistics, Federal University of Paraíba, Cidade Universitária s/n CEP 58.051-900 João Pessoa – PB Brazil, [email protected] 2 Liliane dos Santos Machado, Department of Informatics, Federal University of Paraíba, Cidade Universitária s/n CEP 58.051-900 João Pessoa – PB Brazil, [email protected] Abstract  This paper presents a methodology of evaluation of training executed by Web. It is an intelligent evaluation procedure that allows to classify a trainee learning using a fuzzy rule based expert system. The methodology proposed uses statistical measures, statistical models and appropriate statistical hypothesis testing to compare statistical measures and model parameters. Fuzzy measures, statistical measures and parameters and probability of significance from statistical testing of hypothesis are used as input for a fuzzy rule based expert system to construct a profile for trainee. Fuzzy rules are modeled by membership functions according to two specifications: sequence of tasks executed by experts and statistical measures from execution of correct procedures according to experts.