Abstract:
Las técnicas de EDM (Educational Data Mining) son una herramienta la cual consiste en la implementación de técnicas y algoritmos de minería de datos en bases de datos educativas, la cual hoy en día ha tenido un crecimiento casi exponencial en cuanto a estudios con el propósito de predecir problemas... Ver más
Educational Data Mining (EDM) techniques are a tool that consists of implementing data mining techniques and algorithms in educational databases. Today, there has been an almost exponential growth in studies with the purpose of predicting problems in education such as academic performance, school failure, or early dropout. In addition, they provide a variety of methodologies, which indicates that, on the one hand, there are endless possibilities to be able to test and study different forms of application; and on the other hand, the lack of a consensus on this. However, as far as the method is concerned, practically all studies carry out the same phases: data extraction, which can be done by means of questionnaires or through computer programs on online platforms; data cleaning and preparation, where the attributes will be transformed and, if irrelevant, eliminated; the implementation of an algorithm to try to predict the variable to be studied and, finally, the prediction results are analyzed. In this master’s thesis, the academic performance of two datasets corresponding to schools in Portugal and universities in Bolivia will be predicted. To do this, 3 algorithms have been used (Decision Tree, AdaBoost and Neural Networks) where they have been compared with each other. Quite high accuracies have been obtained as results, without distinction between algorithms. In conclusion, the 3 algorithms used can be used to predict academic performance
|