Student Classification Based on Cognitive Abilities and Predicting Learning Performances Using Machine Learning Models

Abstract

Education is the vital parameter of the country for development in divergent areas like cultivation, economic, political, health, and so on. Any educational Institute’s (universities, colleges, schools) main goal is to increase the student’s learning capabilities and their skills for their full contribution towards society. In these days, “student’s learning process and skill development” research topic requires much-needed attention for the betterment of the society. The student’s performance depends on his/her learning ability and is influenced by many factors. In this paper, we analyze the different categories of student’s leanings that are very fast, fast, moderate, and slow. For this, we conducted the training and tests and use the features likeability, knowledge level, reasoning, and core subject abilities for the 313 engineering students in AITAM, Tekkali, affiliated to JNTUK, India from 2017 to 2019. We gathered information about the personal, academic, cognitive level, and demographic data of students. In this experiment, we are conducting statistical analysis as well as classification of students into 4 types of learners and applying the different Machine Learning (ML) techniques and choose the best ML algorithm for predicting students learning rates. This leads to conducting remedial classes with new teaching methods for moderate and slow learning students. The proposed paper accommodates the individual differences of the learners in terms of knowledge level, learning preferences, cognitive abilities, etc. For this, we apply 5 ML algorithms that are Naive Bayes, Classification Trees (CTs), k-NN, C4.5, and SVM. As per ML analysis, the k-Nearest Neighborhood (k-NN) algorithm is more efficient than other algorithms where the accuracy and prediction values are nearer to 100%.