Identification of Implicit Subject Categories Responsible for Academic Test Scores Using Multiple Linear Regression

Abstract

AbstractIn this work, multiple linear regression technique is employed to study the implicit relation between fundamental categories of a academic topic to the performance of the student. Such a problem is important in the development of personalised learning system. Initially a matrix representation of test questions with its fundamental categories is discussed. Subsequently the responses of the test questions are related to the fundamental categories using a system of linear equations. Multiple Linear Regression (MLR) technique is employed to study the effect of each fundamental categories on the test scores. A synthesised data-set of known probability distributions is used during evaluations. It has been found that the MLR using F-Statistics can identify the complexity of fundamental categories and useful for recommending questions in a personalised learning system.KeywordsMultiple linear regressionPersonalised learning systemF-statistic