engleski

Predicting student satisfaction with courses based on log data from a virtual learning environment – a neural network and classification tree model

In academic institutions students‘ course satisfaction has become an important issue over the years and is recognized as a support in ensuring effective and quality education as well as in enhancing students‘ studying experience. The aim of this paper is to investigate whether there is a connection between students‘ course log data in virtual learning environment (VLE) such as Moodle and students‘ course satisfaction. Furthermore, the paper explores whether is it possible to develop a successful classification model in order to predict student‘s course satisfaction based on their course log data. The research was conducted at the Faculty of Teacher Education in Osijek at Croatian University and included the analysis of log data and course satisfaction on the sample of third and fourth year students. For each study year log data within one VLE course and satisfaction with that course were obtained. Multilayer Perceptron (MLP) with different activation functions and Radial Basis Function (RBF) neural networks as well as classification tree models were developed, trained and tested in order to classify students into one of two categories of course satisfaction. For the purpose of model comparison, classification accuracy, type I and type II errors and input variable importance were used. The results indicate that MLP neural network model provides the highest average classification accuracy, although t-test of the difference in proportions showed that the difference in performance between the compared models is not statistically significant on the level of significance 0.05.

classification ; neural networks ; classification tree ; course satisfaction ; log data

nije evidentirano