Machine learning-driven development of a behaviour-based student classification system (SCS-B) for enhanced educational analytics.

In educational data mining, educational data extraction and analysis, as well as learning analytics, it is very significant to evaluate the student's performance in various aspects for enhancing the educational standards. The growth of students and achievement gap are assumed as the key maters for several educational institutes and universities worldwide. Hence, the educational sectors invest importantly in things to know the good and poor performances of students in attaining better results. In this regard, several methodologies are derived for predicting the student performances for intimating the concern panel to intervene earlier to enhance the overall results. With that note, this paper involves in developing behaviour based student classification system (SCS-B), using machine learning technique. The model collects the student data set based on some questionnaires, providing significant to certain features of student academics and behaviours. Initially, data pre-processing is done with singular value decomposition, performs outlier detection and dimensionality reduction. Following, training process is carried out with genetic algorithm for avoiding local minimum. The findings indicate that the classifier yields superior classification accuracy while requiring minimal processing time for handling extensive student data.