Classifying student success from their behavioral pattern in online learning using machine learning approach

Students' academic activity in Learning Management Systems (LMS) is strongly correlated with their academic results in online learning. This research aims to classify student success through student behavior patterns. The machine learning methods used in this research are Recurrent Neural Network (RNN) to predict student behavioral pattern through time-series data activity recorded in LMS UI, and Support Vector Machine (SVM) to determine if students pass online learning courses or not. Classification in SVM incorporates a pass/fail category. This research uses Recursive Feature Elimination-Random Forest in RNN to select academic activity features that affect the online learning process the most. The best RNN model uses hyperparameters: the number of nodes in the input, hidden and output layers, which are 1, 10 and 1 respectively, as well as a learning rate of 0.01, and 500 epochs on 60% training data. The MSE testing values for three most influential student behavioral patterns are 0.19%, 0.34% and 0.23% respectively. Support Vector Machine can classify student success based on LMS student behavior patterns which are the three features of RFE-RF results with an average precision of 91.1% and an average F1-score of 93.9%, the training data with a proportion of 60%.