Predicting engineering students' optimal group size using socio-educational features

Abstract

Socio-educational background plays an influential role in the success of a studentś engineering schooling. These socio-educational backgrounds are of more diverse nature in developing countries like Bangladesh. The fact that, tertiary education is given in a foreign language adds another dimension to challenge of imparting a successful engineering education. If the students could be grouped according to their socio-educational features, then it would have been easier to anticipate the needs of students coming from diverse backgrounds. In this work, we classify the students (N=237) of the department of Computer Science and Engineering of a university in the Bangladeshi capital of Dhaka based on their socio-educational features using K-means clustering and then propose a classifier that could work as a predictor that could work as a predictor for predicting student needs coming from different backgrounds.