Disease diagnosis and control by reinforcement learning techniques: a systematic literature review
Abstract Disease diagnosis and control is one of the widely recognized and pursued research challenges in the field of reinforcement learning (RL). Early diagnosis and prevention of critical diseases are the major problems, and addressing these can help patients to save lives. Many researchers desig...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-06-01
|
| Series: | Discover Computing |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s10791-025-09539-9 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Disease diagnosis and control is one of the widely recognized and pursued research challenges in the field of reinforcement learning (RL). Early diagnosis and prevention of critical diseases are the major problems, and addressing these can help patients to save lives. Many researchers designed and employed models based on RL to diagnose and control diseases in an early stage. An in-depth systematic literature review (SLR) is presented to find state of the art in the field of RL techniques for the diagnosis and control of critical diseases from 2019 to 2023. In this SLR, 97 research papers were collected from the Scopus database to answer the research questions based on set inclusion criteria, exclusion criteria, study selection, and execution steps. These studies are discussed in detail to discover the recent trends, key findings and research challenges that stimulate the area of RL for disease diagnosis and control. Advanced reinforcement techniques including deep reinforcement, multi-agent reinforcement, hierarchical RL, inverse RL, and federated learning offer substantial promise for improving disease diagnosis and control as healthcare increasingly relies on data-driven and machine learning technologies. These techniques outperformed in the diagnosis and control of critical diseases like thorax disease, urinary disease, epidemic control, cancer, cardiovascular, and heart diseases. However, practical implementation of the model(s) using these techniques will face issues such as data imbalance, data scarcity, safety, and others. Also, the identified limitations and potential future work can be valuable for discerning research trends and developing new reinforcement models. |
|---|---|
| ISSN: | 2948-2992 |