Proyecto de detección de patrones de participación empleando minería de datos en el entorno virtual de aplicaciones web de la ESPOCH, para predecir estudiantes exitosos

Abstract

This research defines a prediction model based on the algorithm for, data mining Multilayer Perceptron was selected by the ROC curve; in this way it is used, historical data of Virtual Learning Environment (VLE) of the subject matter: Web Application in ninth semester at the school of Virtual Platform MOODLE of ESPOCH and the Academic System; in the same way the pre-processing was performed in the PostgreSQL engine two Data Mart were created; using a start-relational dimensional model, where the dimension tables and made with their respective measures were created, the result of the analysis is the information contained in the fact table, submitted to weka and R to identify patterns and its subsequent prediction. Teaching computer data School of Systems Engineering of the current semester, it was submitted to the prediction model to evaluate the model. The algorithm provides of 90% certainty, with this prediction the null hypothesis was accepted.