Student attention evaluation system using machine learning for decision making

Abstract

The student attention evaluation is a very important feature that has a high influence on student results, so it’s necessary to consider. This paper focuses on the evaluation of the student attention level using a machine learning categorization model using several machine learning techniques. The Support Vector Machine, Nearest Neighbor, Naive Bayes, Neural Networks and Random Forest algorithms are applied to model an intelligent system which will evaluate the attention level of the students. Thirteen important features such as alltime, percApp (time in the application), age, grade, behavior biometrics of keyboard (kdt - key down time, tbk – time between keys) and behavior biometrics of mouse (cd - click duration, mv – mouse velocity, ma – mouse acceleration, ddc – duration distance clicks, dplbc – distance point to line between clicks, dbc – distance between clicks, and tbc – time between clicks) are taken for training and testing. Above mentioned machine learning techniques are compared in terms of accuracy rate.