Students Dropout Prediction Based on their Previous Score Using Various Machine Learning Algorithms

Abstract

Student dropout prediction is an important task in educational data analytics that can be used to intervene early and achieve improved academic outcomes. This paper introduces an effective prediction model using the past performance of the students in seven subjects, which include Mathematics, Physics, Chemistry, Biology, English, Social Science and Computer Science. To ensure uniformity and improve the model performance, preprocessing of the data is done through data cleaning, normalization and feature scaling. A number of baseline ML models are implemented such as LR, DT, SVM to be compared. A multi-layer neural network model is suggested to improve predictive power, which is expected to reflect intricate nonlinear relationships among scores of the subjects. Standard measures of performance of all models include accuracy, precision, recall and F1-score. Through experimentation, it is observed that the proposed neural network model will attain a peak accuracy of 89.60% in comparison to the baseline models. The results demonstrate the efficiency of deep learning methods to properly recognize at-risk students and provide educational intervention in a timely manner.

Introduction

The text discusses a neural network-based approach for predicting student dropout using educational data mining. Student dropout prediction is important because it helps identify at-risk students early and allows timely interventions to improve academic performance. While traditional machine learning models such as logistic regression and decision trees have been used, they often struggle with complex, nonlinear relationships in academic data and require extensive manual feature engineering.

To overcome these limitations, the study proposes a neural network model that analyzes student performance across multiple subjects (such as mathematics, physics, chemistry, biology, English, social science, and computer science) to predict whether a student will pass or fail. Neural networks are chosen because they can automatically learn complex patterns and interactions between features, improving prediction accuracy and reliability.

The literature review shows that earlier studies successfully applied machine learning for dropout prediction, but many models lack adaptability, struggle with complex datasets, or fail to generalize well. Recent research suggests that deep learning and hybrid models offer better performance due to their ability to handle high-dimensional and nonlinear data.

The methodology includes data collection from around 1000 student records, data preprocessing (handling missing values, outliers, normalization, and encoding), and training a multi-layer neural network. The system processes cleaned data to produce final predictions and classify students into pass or fail categories.

Conclusion

In this research, a set of data regarding academic performance of various participants was used to create a neural network-based predictor of student dropout. The authors compared the effectiveness of the proposed model to traditional ML methods, which are LR, DT, and SVM, where subject-wise scores are used as input data. The results of the experiment proved the ability of the proposed neural network model to classify students as pass or fail correctly by demonstrating that it outperforms in all the assessment parameters. The model is able to capture the complex and nonlinear interactions between academic variables, which traditional models often fail to capture, as reflected in the higher accuracy, precision, recall, and F1-score. Moreover, the equal level of performance according to a wide range of measures underlines the reliability and stability of the proposed approach. The results highlight how crucial it is to use advanced learning strategies in educational data analysis in order to facilitate the early detection of pupils who are at danger. These prediction algorithms would assist educators and institutions to make informed decisions and implement timely interventions in order to enhance the student success rates and reduce the rate of dropouts. To further improve the performance of prediction, future studies can focus on the further development of the proposed model by incorporating other features such as attendance history, behavioral profiles, and socioeconomic factors. More complex deep learning structures like recurrent neural networks and hybrid systems can be added to establish temporal relationships in student performance data. Moreover, the implementation of the model in the real-time educational systems can help to support the proactive approach to intervention, as it will help to monitor and predict dynamically. Future research could also explore model interpretability measures in order to provide instructors with useful information.

References

[1] G. W. Dekker, M. Pechenizkiy, and J. M. Vleeshouwers, “Predicting students drop out: A case study,” in Proc. 2nd Int. Conf. Educational Data Mining (EDM), Cordoba, Spain, Jul. 2009, pp. 41–50. [2] C. Márquez-Vera, A. Cano, C. Romero, A. Y. M. Noaman, H. Mousa Fardoun, and S. Ventura, “Early dropout prediction using data mining: A case study with high school students,” Expert Systems, vol. 33, no. 1, pp. 107–124, 2016. [3] M. Nagy and R. Molontay, “Predicting dropout in higher education based on secondary school performance,” in 2018 IEEE 22nd Int. Conf. Intelligent Engineering Systems (INES), 2018, pp. 389–394. [4] V. Realinho, J. Machado, L. Baptista, and M. V. Martins, “Predicting student dropout and academic success,” Data, vol. 7, no. 11, p. 146, 2022. [5] S. Lee and J. Y. Chung, “The machine learning-based dropout early warning system for improving the performance of dropout prediction,” Applied Sciences, vol. 9, no. 15, p. 3093, 2019. [6] F. Del Bonifro, M. Gabbrielli, G. Lisanti, and S. P. Zingaro, “Student dropout prediction,” in Proc. Int. Conf. Artificial Intelligence in Education (AIED), Cham, Switzerland: Springer, Jun. 2020, pp. 129–140. [7] W. Tenpipat and K. Akkarajitsakul, “Student dropout prediction: A KMUTT case study,” in 2020 1st Int. Conf. Big Data Analytics and Practices (IBDAP), 2020, pp. 1–5. [8] V. Christou et al., “Performance and early drop prediction for higher education students using machine learning,” Expert Systems with Applications, vol. 225, p. 120079, 2023. [9] Z. Song, S. H. Sung, D. M. Park, and B. K. Park, “All-year dropout prediction modeling and analysis for university students,” Applied Sciences, vol. 13, no. 2, p. 1143, 2023. [10] S. Kim, E. Choi, Y. K. Jun, and S. Lee, “Student dropout prediction for university with high precision and recall,” Applied Sciences, vol. 13, no. 10, p. 6275, 2023.

Copyright

Copyright © 2026 K Sai Shri, D. Anil Kumar . This 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.