University Information System’s Impact on Academic Performance: A Comprehensive Logistic Regression Analysis with Principal Component Analysis and Performance Metrics

Selim, Aybeyan and Ali, Ilker and Ristevski, Blagoj University Information System’s Impact on Academic Performance: A Comprehensive Logistic Regression Analysis with Principal Component Analysis and Performance Metrics. TEM JOURNAL - Technology, Education, Management, Informatics. ISSN 2217-8309

Text TEMJournalMay2024_1589_1598-7.pdf Download (787kB)

Abstract

This paper comprehensively analyzes data mining techniques and performance metrics applied to a logistic regression model. Principal Component Analysis (PCA) was utilized to diminish the complexity of high-dimensional data, enabling clearer visualization and examination of intricate relationships among variables. The logistic regression model demonstrated commendable performance on both test and train sets, as evidenced by high values of accuracy, precision, recall, ROC AUC, and F1 Score were observed. The provided confusion matrices offered detailed insights into the model's accuracy in classifying positive and negative instances. Concerning our hypotheses, we found no significant relationship between gender and academic performance, supported by a highly significant p-value of 0.0 and a weak positive correlation coefficient of 0.0847. However, we noticed a strong positive correlation of 0.99 between gender and exam characteristics, although it has not
reached statistical significance for a p-value of 0.281.
Our research contributes valuable insights into data analysis, model evaluation, and the interplay between
variables.

Item Type:	Article
Subjects:	Scientific Fields (Frascati) > Natural sciences > Computer and information sciences Scientific Fields (Frascati) > Engineering and Technology > Electrical engineering, electronic engineering,information engineering
Divisions:	Faculty of Information and Communication Technologies
Depositing User:	Prof. d-r. Blagoj Ristevski
Date Deposited:	23 Oct 2024 17:20
Last Modified:	23 Oct 2024 17:20
URI:	https://eprints.uklo.edu.mk/id/eprint/10383

Actions (login required)

View Item