Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM69290
Volume:
10 |
11 |
Nov 2026
Pages:
1-6
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Master of Computer Applications
Nehru College of Management
Master of Computer Applications
Nehru College of Management
Identifying students who may experience academic difficulty at an early stage is an important task for educational institutions. Traditional review often depends on periodic examinations and manual interpretation of marks, attendance and student-related information. This paper presents EduPredict, an explainable artificial intelligence-based student performance prediction system designed to estimate a student’s academic outcome as Pass or Fail and provide information that helps educators understand the prediction. The project uses a student-score dataset containing subject marks and selected contextual attributes, including gender, part-time employment, extracurricular participation, career aspiration, absences and weekly self-study hours. Direct identifiers such as student ID, first name, last name and email are excluded from model inputs. Categorical attributes are converted into numerical representations, and a binary target is generated using the project rule: a student is labelled Pass when the score is at least 40 in six or more of the seven subjects; otherwise, the label is Fail. A Random Forest classifier with 100 decision trees is trained using an 80:20 train–test split. The trained model and feature-column information are saved for use by a Streamlit application. SHAP is used to explain feature contributions for individual predictions. The application supports single-student prediction, batch CSV prediction, visual summaries, and storage of user and prediction records through SQLite. The current training script creates a test partition but does not calculate or print evaluation metrics; therefore, no numerical accuracy result is claimed in this paper. The system is intended as a decision-support prototype that can help educators review students who may need assistance, rather than as a replacement for professional judgement.
KEYWORDS
This paper presented EduPredict, an explainable AI-based student performance prediction prototype that estimates Pass or Fail outcomes from academic and selected contextual attributes. The system uses a Random Forest classifier configured with 100 trees, an 80:20 train–test split, saved preprocessing artifacts and SHAP-based feature explanations. A Streamlit interface supports single-student and batch prediction, while SQLite stores user accounts and prediction history. The project demonstrates an integrated workflow for educational data processing, machine-learning inference and explainable output. However, the current training script does not report test-set metrics, and the README accuracy claim requires independent verification. Quantitative evaluation is therefore an essential next step. Future work includes calculating accuracy, precision, recall, F1-score and a confusion matrix; applying cross-validation; checking class imbalance; validating uploaded data; strengthening account and administrator security; and conducting fairness and privacy reviews. Further work may also explore model comparison and additional explanation views after the current implementation has been evaluated. With these improvements, EduPredict can become a more dependable tool for helping educators identify students who may benefit from timely academic support.
[1] L. Breiman, “Random Forests,” Machine Learning, vol. 45, pp. 5–32, 2001. [2] S. M. Lundberg and S.-I. Lee, “A Unified Approach to Interpreting Model Predictions,” in Advances in Neural Information Processing Systems, vol. 30, 2017. [3] F. Pedregosa et al., “Scikit-learn: Machine Learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011. [4] W. McKinney, “Data Structures for Statistical Computing in Python,” in Proceedings of the 9th Python in Science Conference, pp. 56–61, 2010. [5] Streamlit, “Streamlit Documentation.” Available:https://docs.streamlit.io/. Accessed: Nov. 19, 2026. [6] SHAP, “SHAP Documentation.” Available:https://shap.readthedocs.io/. Accessed: Nov. 19, 2026. [7] Scikit-learn, “RandomForestClassifier Documentation.”Available: https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.Rando mForestClassifier.html. Accessed: Nov. 19, 2026. 8] C. Romero and S. Ventura, “Educational data mining: A review of the state of the art,” IEEE Trans. Syst., Man, Cybern. C, vol. 40, no. 6, pp. 601–618, Nov. 2010. [9] R. S. Baker and K. Yacef, “The state of educational data mining in 2009: A review and future visions,” J. Educ. Data Mining, vol. 1, no. 1, pp. 3–17, 2009. [10] A. Peña-Ayala, “Educational data mining: A survey and a data mining-based analysis of recent works,” Expert Syst. Appl., vol. 41, no. 4, pp. 1432–1462, 2014.
Publication Date
11/20/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
0
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...