Student Performance Prediction Using Machine Learning

  • Unique Paper ID: 211376
  • PageNo: 24-31
  • Abstract:
  • Predicting student academic performance can help educators identify learners who may benefit from timely academic support. This study compares five supervised machine-learning classifiers—Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors (KNN), and Support Vector Machine (SVM)—using the mathematics portion of the UCI Student Performance dataset. The target variable, final grade (G3), is grouped into three classes: Low (0–9), Average (10–13), and High (14–20). Two experiments are conducted: Experiment A includes the earlier period grades G1 and G2, while Experiment B excludes them to examine prediction using other available attributes. The supplied experimental output reports 395 records and a class distribution of 130 Low, 165 Average, and 100 High instances. In Experiment A, Random Forest achieved the highest test accuracy (86.08%) and Macro F1-score (0.8684), with a cross-validation Macro F1 mean of 0.8479. In Experiment B, SVM achieved the highest test accuracy (53.16%) and Macro F1-score (0.5222), while Logistic Regression had a test Macro F1-score of 0.4813. Overall, removing G1 and G2 substantially reduced performance, indicating that earlier grades contain strong predictive information in this dataset. Permutation-importance output identified failures as the most positively important feature in Experiment B, although most reported importance values were close to zero or negative and should be interpreted cautiously. The results are specific to this dataset and require validation on independent student populations before practical deployment.

Copyright & License

Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

BibTeX

@article{211376,
        author = {Anuja Medhe and Sakshi Mahale and Smita Ghorpade},
        title = {Student Performance Prediction Using Machine Learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {no},
        pages = {24-31},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=211376},
        abstract = {Predicting student academic performance can help educators identify learners who may benefit from timely academic support. This study compares five supervised machine-learning classifiers—Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors (KNN), and Support Vector Machine (SVM)—using the mathematics portion of the UCI Student Performance dataset. The target variable, final grade (G3), is grouped into three classes: Low (0–9), Average (10–13), and High (14–20). Two experiments are conducted: Experiment A includes the earlier period grades G1 and G2, while Experiment B excludes them to examine prediction using other available attributes. The supplied experimental output reports 395 records and a class distribution of 130 Low, 165 Average, and 100 High instances. In Experiment A, Random Forest achieved the highest test accuracy (86.08%) and Macro F1-score (0.8684), with a cross-validation Macro F1 mean of 0.8479. In Experiment B, SVM achieved the highest test accuracy (53.16%) and Macro F1-score (0.5222), while Logistic Regression had a test Macro F1-score of 0.4813. Overall, removing G1 and G2 substantially reduced performance, indicating that earlier grades contain strong predictive information in this dataset. Permutation-importance output identified failures as the most positively important feature in Experiment B, although most reported importance values were close to zero or negative and should be interpreted cautiously. The results are specific to this dataset and require validation on independent student populations before practical deployment.},
        keywords = {student performance prediction, educational data mining, machine learning, classification, Random Forest, SVM, Macro F1-score, feature importance},
        month = {November},
        }

Cite This Article

Medhe, A., & Mahale, S., & Ghorpade, S. (2026). Student Performance Prediction Using Machine Learning. International Journal of Innovative Research in Technology (IJIRT), 24–31.

Related Articles