Article: Analysis of factors affecting college students' academic performance based on linear regression Journal: International Journal of Computational Systems Engineering (IJCSYSE) 2026 Vol.10 No.8 pp.1 - 13 Abstract: The article examines determinants of college students' academic performance (AP) by using an extended linear regression model (ELRM). Unlike traditional linear regression models (TLRMs), which fail to properly account for the time-varying effects and interactions of dynamic behavioural determinants, the ELRM incorporates time series characteristics to model the temporal and interactive effects on AP. Based on data spanning several semesters, the study finds that short-term learning habits (LHs), such as cramming, have a greater immediate contribution to grades compared to long-term studying habits. The ELRM performs better than conventional models with a considerably lower mean squared error (MSE = 0.81) and a better coefficient of determination (R<SUP align="right"><SMALL>2</SMALL></SUP> = 0.985), demonstrating its enhanced predictive ability and stability across semesters. The study highlights the value of temporal determinants and interaction effects in explaining the dynamics of AP, providing recommendations for enhancing educational interventions and policy. Inderscience Publishers - linking academia, business and industry through research

Title: Analysis of factors affecting college students' academic performance based on linear regression

Authors: Ru Huang

Addresses: Academic Affairs Office, Guilin University of Aerospace Technology, Jinji Road, Qixing District, Guilin, Guangxi, China

Abstract: The article examines determinants of college students' academic performance (AP) by using an extended linear regression model (ELRM). Unlike traditional linear regression models (TLRMs), which fail to properly account for the time-varying effects and interactions of dynamic behavioural determinants, the ELRM incorporates time series characteristics to model the temporal and interactive effects on AP. Based on data spanning several semesters, the study finds that short-term learning habits (LHs), such as cramming, have a greater immediate contribution to grades compared to long-term studying habits. The ELRM performs better than conventional models with a considerably lower mean squared error (MSE = 0.81) and a better coefficient of determination (R² = 0.985), demonstrating its enhanced predictive ability and stability across semesters. The study highlights the value of temporal determinants and interaction effects in explaining the dynamics of AP, providing recommendations for enhancing educational interventions and policy.

Keywords: academic performance; extended linear regression; study habits; time series analysis; learning behaviour; cramming; dynamic interaction; data mining.

DOI: 10.1504/IJCSYSE.2026.153273

International Journal of Computational Systems Engineering, 2026 Vol.10 No.8, pp.1 - 13

Received: 13 Aug 2025
Accepted: 22 Jan 2026
Published online: 29 Apr 2026 *

Title: Analysis of factors affecting college students' academic performance based on linear regression

Keep up-to-date