A Comprehensive Review of Medicine Recommendation Systems Using Machine Learning and Deep Learning Techniques

Abstract

There is a growing need for personalized healthcare services. This has led to the development of systems that help patients and healthcare professionals choose the right medicines. Old methods for recommending medicines relied on knowledge, rules and guidelines. With advancements in Artificial Intelligence Machine Learning, Deep Learning, Natural Language Processing and Semantic Web technologies healthcare recommendations have become more accurate, scalable and personalized. This review looks at existing medicine recommendation systems. We examine how recommendation methods have changed over time. We go from methods like collaborative filtering and content-based filtering to modern methods like machine learning, deep learning and sentiment analysis. We review algorithms like Decision Trees, Random Forests and Support Vector Machines. We also look at Recurrent Neural Networks, Long Short-Term Memory and Bidirectional Gated Recurrent Units. Ontology-based clinical decision support systems and Natural Language Processing-driven drug recommendation models are also analyzed. These models use reviews and healthcare text data.We discuss used datasets, evaluation metrics and system architectures. We also talk about challenges and emerging trends in healthcare recommendation systems. Medicine recommendation systems and Machine Learning are areas of research. Deep Learning and Natural Language Processing-based approaches are effective in disease prediction and personalized medicine recommendation tasks. Semantic Web technologies improve explainability and clinical safety. Current research gaps and future directions are highlighted. These include AI, knowledge graphs, federated learning and transformer-based healthcare models. These will guide developments, in intelligent medicine recommendation systems.

Introduction

The text reviews the development of intelligent medicine recommendation systems that use Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL), Natural Language Processing (NLP), and semantic technologies to assist in disease prediction and medication selection.

The growing complexity of healthcare and the large number of available medicines have created a need for automated systems that can support healthcare professionals and patients in making treatment decisions. Traditional diagnosis and medication selection rely heavily on doctors’ expertise and patient records, which can be time-consuming and prone to human error. Medicine recommendation systems help address these challenges by analyzing symptoms, medical history, and healthcare data to suggest suitable treatments.

Early recommendation systems adopted techniques from e-commerce, such as content-based filtering, collaborative filtering, and hybrid recommendation methods. Over time, more advanced approaches emerged due to the healthcare sector’s strict requirements for safety, accuracy, and reliability. Machine learning models such as Decision Trees, Random Forests, Support Vector Machines (SVM), Naïve Bayes, and Neural Networks have shown strong performance in predicting diseases and recommending medicines based on patient symptoms.

Recent advancements include the use of Natural Language Processing (NLP) to analyze unstructured medical information such as clinical notes, patient reviews, and symptom descriptions. Deep learning models, particularly Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), and Bidirectional GRU (BiGRU), have improved disease prediction by effectively understanding sequential symptom patterns and medical text.

Another important development is the use of ontology and semantic web technologies, which provide structured medical knowledge and reasoning capabilities. Systems such as GalenOWL and Panacea improve recommendation explainability, detect drug-drug interactions, and enhance prescription safety.

The literature review highlights several notable studies:

• Hybrid recommender systems improved recommendation accuracy.
• Ontology-based systems enhanced explainability and drug interaction detection.
• SVM-based systems achieved high medicine recommendation accuracy.
• Machine learning-based systems successfully predicted diseases and suggested medications.
• Naïve Bayes models achieved prediction accuracies of up to 98.12%.
• Random Forest-based systems achieved around 94.2% accuracy.
• Recent personalized medicine recommendation systems incorporate patient lifestyle, medical history, diet, and exercise suggestions.

The paper identifies several ongoing challenges:

• Data privacy and security concerns.
• Limited explainability of AI models.
• Maintaining up-to-date medical knowledge bases.
• Accurate detection of drug interactions and adverse effects.
• Need for clinical validation in real healthcare environments.
• Ensuring personalized and safe recommendations.

The healthcare recommendation system generally consists of key modules:

1. Data Collection Module – gathers symptoms, patient records, medical databases, and drug information.
2. Data Preprocessing Module – cleans and prepares healthcare data.
3. Disease Prediction Module – predicts diseases using AI and ML models.
4. Medicine Recommendation Module – suggests suitable medications.
5. Knowledge and Interaction Analysis Module – identifies drug interactions and safety concerns.

References

[1] Y. Bao and X. Jiang, “An Intelligent Medicine Recommender System Framework,” 2016 IEEE ICIEA, 2016. DOI: 10.1109/ICIEA.2016.7603801 [2] R. C. Chen, J. Y. Chiu, and C. T. Bau, “The Recommendation of Medicines Based on Multiple Criteria Decision Making and Domain Ontology,” ICMLC, 2011. DOI: 10.1109/ICMLC.2011.6016685 [3] R. C. Chen, Y. H. Huang, C. T. Bau, and S. M. Chen, “A Recommendation System Based on Domain Ontology and SWRL for Anti-Diabetic Drugs Selection,” Expert Systems with Applications, 2012. DOI: 10.1016/j.eswa.2011.09.061 [4] C. Doulaverakis, G. Nikolaidis, A. Kleontas, and I. Kompatsiaris, “GalenOWL: Ontology-Based Drug Recommendations Discovery,” Journal of Biomedical Semantics, 2012. DOI: 10.1186/2041-1480-3-14 [5] C. Doulaverakis, G. Nikolaidis, A. Kleontas, and I. Kompatsiaris, “Panacea: A Semantic-Enabled Drug Recommendations Discovery Framework,” Journal of Biomedical Semantics, 2014. DOI: 10.1186/2041-1480-5-13 [6] A. A. Hamed, R. Roose, M. Branicki, and A. Rubin, “T-Recs: Time-Aware Twitter-Based Drug Recommender System,” IEEE/ACM ASONAM, 2012. DOI: 10.1109/ASONAM.2012.178 [7] B. Stark, C. Knahl, M. Aydin, and K. Elish, “A Literature Review on Medicine Recommender Systems,” IJACSA, 2019. DOI: 10.14569/IJACSA.2019.0100802 [8] S. Garg, “Drug Recommendation System Based on Sentiment Analysis of Drug Reviews Using Machine Learning,” IEEE Confluence, 2021. DOI: 10.1109/Confluence51648.2021.9377188 [9] P. Keikhosrokiani, K. Balasubramaniam, and M. Isomursu, “Drug Recommendation System for Healthcare Professionals’ Decision-Making Using Opinion Mining and Machine Learning,” Springer, 2024. DOI: 10.1007/978-3-031-59091-7_15 [10] M. Zomorodi et al., “RECOMED: A Comprehensive Pharmaceutical Recommendation System,” Artificial Intelligence in Medicine, 2024. DOI: 10.1016/j.artmed.2024.102981 [11] M. Sajde, H. Malek, and M. Mohsenzadeh, “RecoMed: A Knowledge-Aware Recommender System for Hypertension Medications,” Informatics in Medicine Unlocked, 2022. DOI: 10.1016/j.imu.2022.100950 [12] J. P. Gupta, A. Singh, and R. K. Kumar, “A Computer-Based Disease Prediction and Medicine Recommendation System Using Machine Learning Approach,” IJARET, 2021. DOI: 10.34218/IJARET.12.3.2021.062 [13] N. V. Shaik et al., “Medicine Recommendation System: Health Harbour,” WJARR, 2025. DOI: 10.30574/wjarr.2025.25.2.0382 [14] K. P. Swain et al., “Leveraging Machine Learning and Patient Reviews for Developing a Drug Recommendation System to Reduce Medical Errors,” Open Bioinformatics Journal, 2024. DOI: 10.2174/0118750362291402240621044046 [15] K. Tiwari and D. K. Singh, “Machine Learning-Based Recommendation System for Disease-Drug Material and Adverse Drug Reaction: Comparative Review,” Materials Today: Proceedings, 2022. DOI: 10.1016/j.matpr.2021.05.404 [16] L. Schmidt-Thieme, “Compound Classification Models for Recommender Systems,” IEEE ICDM, 2005. DOI: 10.1109/ICDM.2005.46 [17] F. Ricci, L. Rokach, and B. Shapira, Recommender Systems Handbook, Springer, 2015. DOI: 10.1007/978-1-4899-7637-6 [18] C. C. Aggarwal, Recommender Systems: The Textbook, Springer, 2016. DOI: 10.1007/978-3-319-29659-3 [19] P. Lops, M. de Gemmis, and G. Semeraro, “Content-Based Recommender Systems: State of the Art and Trends,” Springer, 2011. DOI: 10.1007/978-0-387-85820-3_3 [20] J. Bobadilla, F. Ortega, A. Hernando, and A. Gutiérrez, “Recommender Systems Survey,” Knowledge-Based Systems, 2013. DOI: 10.1016/j.knosys.2013.03.012 [21] R. Burke, “Hybrid Recommender Systems: Survey and Experiments,” User Modeling and User-Adapted Interaction, 2002. DOI: 10.1023/A:1021240730564 [22] S. Zhang, L. Yao, A. Sun, and Y. Tay, “Deep Learning Based Recommender System: A Survey and New Perspectives,” ACM Computing Surveys, 2019. DOI: 10.1145/3285029 [23] G. Adomavicius and A. Tuzhilin, “Toward the Next Generation of Recommender Systems,” IEEE TKDE, 2005. DOI: 10.1109/TKDE.2005.99 [24] X. Su and T. M. Khoshgoftaar, “A Survey of Collaborative Filtering Techniques,” Advances in Artificial Intelligence, 2009. DOI: 10.1155/2009/421425 [25] Y. Koren, R. Bell, and C. Volinsky, “Matrix Factorization Techniques for Recommender Systems,” Computer, 2009. DOI: 10.1109/MC.2009.263 [26] T. Chen and C. Guestrin, “XGBoost: A Scalable Tree Boosting System,” ACM KDD, 2016. DOI: 10.1145/2939672.2939785 [27] J. H. Friedman, “Greedy Function Approximation: A Gradient Boosting Machine,” Annals of Statistics, 2001. DOI: 10.1214/aos/1013203451

Copyright

Copyright © 2026 Abhishek Kumar Yadav, Kumar Bibhuti Bhushan Singh, Anita Pal, Neha Singh, Shivam Dixit. 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.