Clustering students based on Student's performance - A partial least squares path modeling (PLS-PM) study

Abstract

Monitoring and evaluation of student's performance is an important task in learning process. The analysis of relation between student performance and other variables in education setting often can be useful in identifying influential factors on performance. This in turn may help to categorize students and provide more focus towards improving their performance. In this paper, the student performance is analyzed based on its relationship with medium of study. More precisely, among 10th grade students (Karnataka SSLC Board) the relationship between performance in their main subjects (Mathematics, Science and Social Science) and their medium of study is analyzed. The analysis is based on the research hypothesis: "The performance in the language of medium of study influences the performance in main subjects". This paper presents a theoretical model that shows how the data related to language of medium of study and the main subjects are modeled and evaluated. The results from the experiments indicate a strong positive plausible causal relationship between medium of study and performance. Clusters of students are identified based on the degree of influence. The contributions in this paper are two fold. 1) Novel clustering algorithm in PLS-PM and 2) Identification of plausible causal indicators for students performance and analysis of the same. © 2014 Springer International Publishing Switzerland.