Explainable Deep Learning for Predicting Student Dropout in Kenyan Universities: A Systematic Review of Models, Risk Factors, and Interpretability
Keywords:
Deep learning, Explainable artificial intelligence, SHAP, Student dropout, KenyaAbstract
Student dropout continues to hinder higher education in Kenya, and one of the main defenses against it is identifying at-risk students early. Deep learning outperforms conventional machine learning in predictive power, yet universities remain cautious about adopting it because its decisions are hard to explain. Explainable artificial intelligence tools such as SHAP and LIME are meant to open up these black-box predictions for university officials, but whether they have actually been applied to dropout prediction in Kenya or elsewhere in Africa is still unclear. This review gathered and assessed the available evidence on explainable deep learning for student dropout prediction, with particular attention to what it means for Kenyan universities. Eleven keywords guided the search, returning 1,628 candidate records. After removing duplicates, screening titles and abstracts, and checking eligibility, 26 papers were retained for evaluation. Risk of bias was assessed with PROBAST, and results were synthesized narratively around four research questions covering predictor factors, deep learning architectures, explainable AI application, and relevance to Kenya and Africa. Academic and behavioral variables emerged as the most consistently reported predictors of dropout. Hybrid architectures that pair a deep feature extractor with a second-stage classifier delivered the strongest predictive performance whenever they were compared directly against single-architecture models. Only five of the 26 studies applied SHAP or LIME, and only two studies were conducted in Africa, one of them in Kenya. Deep learning has built a fairly solid evidence base for dropout prediction, but interpretability has not kept up with how complex these models have become, and evidence from Kenya remains scarce. This gap points to both the research that is still needed and the design choices that will shape a planned study built around an explainable deep neural network.
References
Adadi, A., & Berrada, M. (2018). Peeking inside the black-box: A survey on explainable artificial intelligence (XAI). IEEE Access, 6, 52138–52160. https://doi.org/10.1109/ACCESS.2018.2870052
Albugami, S., Almagrabi, H., & Wali, A. (2024). From data to decision: Machine learning and explainable AI in student dropout prediction. Journal of e-Learning and Higher Education, 2024, Article 246301. https://doi.org/10.5171/2024.246301
Alharbi, A., Janarthanan, M., & Midhunchakkaravarthy, D. (2026). SHAP-centric interpretability in educational machine learning: A PRISMA systematic review of SHAP and LIME. Letters in Biomathematics.
Aljawazneh, H., Alqirem, R., & Al-Adwan, N. (2026). A comparative study of machine learning and deep learning algorithms for student dropout prediction in higher education. Al-Zaytoonah University Journal of Business. https://doi.org/10.15849/zjjb.v2i1.16
Altabrawee, H., Ali, O. A. J., & Ajmi, S. Q. (2019). Predicting students' performance using machine learning techniques. Journal of University of Babylon for Pure and Applied Sciences, 27(1), 194–205.
Baker, R. S., & Yacef, K. (2009). The state of educational data mining in 2009: A review and future visions. Journal of Educational Data Mining, 1(1), 3–17. https://doi.org/10.5281/zenodo.3554657
Baranyi, M., Nagy, M., & Molontay, R. (2020). Interpretable deep learning for university dropout prediction. In SIGITE '20: Proceedings of the 21st Annual Conference on Information Technology Education. https://doi.org/10.1145/3368308.3415382
Bean, J. P., & Metzner, B. S. (1985). A conceptual model of nontraditional undergraduate student attrition. Review of Educational Research, 55(4), 485–540 https://doi.org/10.3102/00346543055004485
Bettahi, A., Belouadha, F.-Z., & Harroud, H. (2025). A modular and explainable machine learning pipeline for student dropout prediction in higher education. Algorithms, 18(10), 662. https://doi.org/10.3390/a18100662
Cheng, Y.-H., & Lin, C.-E. (2026). Predicting university student dropout risk using deep learning and ensemble voting mechanism. Engineering Proceedings. https://doi.org/10.3390/engproc2025120066
Chung, J. Y., & Lee, S. (2019). Dropout early warning systems for high school students using machine learning. Children and Youth Services Review, 96, 346–353. https://doi.org/10.1016/j.childyouth.2018.11.030
Commission for University Education. (2024). University statistics 2024/2025 report. https://www.cue.or.ke
Coppo, E. C., Caetano, R. S., de Lima, L. M., & Krohling, R. A. (2022). Student dropout prediction using 1D CNN-LSTM with variational autoencoder oversampling. In 2022 IEEE Latin American Conference on Computational Intelligence (LA-CCI). https://doi.org/10.1109/LA-CCI54402.2022.9981340
Debray, T. P. A., Damen, J. A. A. G., Snell, K. I. E., Ensor, J., Hooft, L., Reitsma, J. B., & Moons, K. G. M. (2017). A guide to systematic review and meta-analysis of prediction model performance. BMJ, 356, i6460. https://doi.org/10.1136/bmj.i6460
Duro, B., Gomes, A., Correia, F. B., Borges, A. R., & Bernardino, J. (2026). Machine learning and deep learning for dropout prediction in higher education: A review. Computers, 15(3), 164.
Gunning, D. (2017). Explainable artificial intelligence (XAI). Defense Advanced Research Projects Agency (DARPA), 2(2), 1.
Guyatt, G. H., Oxman, A. D., Akl, E. A., Kunz, R., Vist, G., Brozek, J., Norris, S., Falck-Ytter, Y., Glasziou, P., deBeer, H., Jaeschke, R., Rind, D., Meerpohl, J., Dahm, P., & Schunemann, H. J. (2011). GRADE guidelines: 1. Introduction, GRADE evidence profiles and summary of findings tables. Journal of Clinical Epidemiology, 64(4), 383–394. https://doi.org/10.1016/j.jclinepi.2010.04.026
Hamal, O., El Koufi, N., Cengiz, K., & Milić, L. (2025). AI-based prediction of student dropout using an enhanced deep learning framework: DeepDropNet model. International Journal of Advances in Soft Computing & Its Applications. https://doi.org/10.15849/IJASCA.251130.06
Ibrahim, A. M., Abonazel, M. R., & Ahmed, A.-H. N. (2026). A proposed hybrid deep learning and ensemble learning model for predicting student dropout in education. Statistics, Optimization & Information Computing. https://doi.org/10.19139/soic-2310-5070-3733
Kendagor, V. J., Chemwa, G., & Cheruyoit, W. (2023). A convolutional neural network framework for predicting student drop out in universities. In 2023 IEEE AFRICON. https://doi.org/10.1109/AFRICON55910.2023.10293554
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. https://doi.org/10.1038/nature14539
Lopez-Muñoz, G. L., Gonzalez Serrano, C., & Sanchez-Ferreira, C. (2025). Student dropout prediction in higher education: A systematic review of machine learning methods and risk factors. Journal of Information Technology Education: Research, 25, 1. https://doi.org/10.28945/5776
Loritee, L. J., Kyalo, D., & Nyaegah, J. O. (2026). Tuition fee levels and their influence on university student participation: Evidence from the University of Nairobi, Kenya. Journal of Education, 6(4), 16–28.
Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30.
Marcolino, M. R., Porto, T. R., Primo, T. T., Targino, R., Ramos, V., Queiroga, E. M., Munoz, R., & Cechinel, C. (2025). Student dropout prediction through machine learning optimization: Insights from Moodle log data. Scientific Reports, 15, 9840. https://doi.org/10.1038/s41598-025-93918-1
Mduma, N., & Machuve, D. (2021). Machine learning model for predicting student dropout: A case of Tanzania, Kenya and Uganda. In 2021 IEEE AFRICON (pp. 1–6). https://doi.org/10.1109/AFRICON51333.2021.9570956
Moons, K. G. M., de Groot, J. A. H., Bouwmeester, W., Vergouwe, Y., Mallett, S., Altman, D. G., Reitsma, J. B., & Collins, G. S. (2014). Critical appraisal and data extraction for systematic reviews of prediction modelling studies: The CHARMS checklist. PLoS Medicine, 11(10), e1001744. https://doi.org/10.1371/journal.pmed.1001744
Nagy, M., & Molontay, R. (2024). Interpretable dropout prediction: Towards XAI-based personalized intervention. International Journal of Artificial Intelligence in Education, 34(2), 274–300. https://doi.org/10.1007/s40593-024-00412-w
Nyawira, L., Musau, O., Adem, A., & Jobunga, E. (2025). Predicting student attrition in Kenyan universities: A comparative analysis of machine learning algorithms. Journal of the Kenya National Commission for UNESCO, 5(2), 1–9.
Ogawa, M. (2024). Massification of tertiary education and its inequality in Kenya: A case study of top students from a rural day secondary school. International Journal of Educational Development, 110, 103137. https://doi.org/10.1016/j.ijedudev.2024.103137
Oreopoulos, P., & Petronijevic, U. (2013). Making college worth it: A review of the returns to higher education. The Future of Children, 23(1), 41–65. https://doi.org/10.1353/foc.2013.0001
Ouzzani, M., Hammady, H., Fedorowicz, Z., & Elmagarmid, A. (2016). Rayyan, a web and mobile app for systematic reviews. Systematic Reviews, 5, 210. https://doi.org/10.1186/s13643-016-0384-4
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hrobjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71
Rajput, V., Singh, A., & Gupta, S. C. (2025). Early detection of student dropout risk via hybrid deep learning and gradient boosting algorithms. In 2025 International Conference on Engineering, Technology & Management (ICETM). https://doi.org/10.1109/ICETM63734.2025.11051847
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). Why should I trust you? Explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1135–1144). https://doi.org/10.1145/2939672.2939778
Romero, C., & Ventura, S. (2020). Educational data mining and learning analytics: An updated survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 10(3), e1355. https://doi.org/10.1002/widm.1355
Sangeetha, K., & Priya, N. S. (2026). Improved GRU-based temporal modeling for robust student dropout classification and risk prediction. In 2026 International Conference on Smart Electronic Devices and Intelligent Systems (ICSEDIS). https://doi.org/10.1109/ICSEDIS68157.2026.11518148
Sarker, I. H. (2021). Machine learning: Algorithms, real-world applications and research directions. SN Computer Science, 2(3), 160. https://doi.org/10.1007/s42979-021-00592-x
Sivasankar, C., Palani, H. K., Fathima, A., Gunasekaran, K., Agoramoorthy, M., & Vasanthi, G. (2026). TERGATE-Warn: Temporal graph attention for equitable educational early warning. In 2026 6th International Conference on Expert Clouds and Applications (ICOECA). https://doi.org/10.1109/ICOECA68095.2026.11485035
Stofile, R., & Enwereji, C. (2025). Investigating the impact of institutional support services on student retention in a university in the Eastern Cape Province of South Africa. http://dx.doi.org/10.48047/jett.2025.16.06.07
Tinto, V. (2006). Research and practice of student retention: What next? Journal of College Student Retention: Research, Theory & Practice, 8(1), 1–19. https://doi.org/10.2190/4YNU-4TMB-22DJ-AN4W
Tinto, V. (2017). Through the eyes of students. Journal of College Student Retention: Research, Theory & Practice, 19(3), 254–269.
Unnati, M., Nakum, Z., & R, P. S. (2026). Hybrid CNN-LSTM model for student dropout risk prediction with interpretable AI. In 2026 International Conference on Computer Networks and Inventive Communication Technologies (ICCNCT). https://doi.org/10.1109/ICCNCT68477.2026.11590425
Wolff, R. F., Moons, K. G. M., Riley, R. D., Whiting, P. F., Westwood, M., Collins, G. S., Reitsma, J. B., Kleijnen, J., & Mallett, S. (2019). PROBAST: A tool to assess the risk of bias and applicability of prediction model studies. Annals of Internal Medicine, 170(1), 51–58. https://doi.org/10.7326/M18-1376
