Melika Ahmadi Bonabi - Research Center for Evidence-Based Medicine, Iranian EBM Centre: A Joanna Briggs Institute Affiliated Group, Tabriz University of Medical Sciences,Tabriz, Iran
Morteza Ghojazadeh - Research Center for Evidence-Based Medicine, Iranian EBM Centre: A Joanna Briggs Institute Affiliated Group, Tabriz University of Medical Sciences,Tabriz, Iran
Razieh Abdolrahmanzadeh - Research Center for Evidence-Based Medicine, Iranian EBM Centre: A Joanna Briggs Institute Affiliated Group, Tabriz University of Medical Sciences,Tabriz, Iran

Background and aims: Nowadays, one of the areas that is expanding and has a wide range ofuses in the medical sciences is machine learning (ML), which has improved this field’s capabilityand effectiveness in several ways. The aim of this paper is to present an overview of the currentML methods utilized in medical sciences and to identify the evidence gaps and features ofML techniques in medicine.Methods: A comprehensive systematic search was conducted in PubMed using the terms “machinelearning”, “deep learning”, “supervised learning”, “unsupervised learning”, “reinforcementlearning”, “healthcare”, “medicine”, and “medical science” as keywords up to March ۲۰۲۳. Theinclusion criteria were English language and full-text available articles. After removing duplicates,two independent reviewers screened and evaluated the investigated results in accordancewith the inclusion criteria. Data obtained from the included studies was extracted by two independentauthors. From each study, data were taken on the type and model of ML employed, themedical scientific field, user data analysis, challenges faced, and directions for future research.Results: We retrieved ۱۳۰۱۴ relevant publications from electronic databases. After a thoroughexamination of the titles and abstracts, ۲۰۸ articles were included. ML is widely implemented inthe medical sciences. Machine learning models, including Artificial Neural Networks (ANNs),Support Vector Machines (SVMs), Regression analysis and Bayesian Networks (BNs) were mostfrequently applied to medical sciences. The classification, prediction, and processing of the enormousamounts of data available in this area is one of the most significant uses of ML nowadays.Conclusion: ML can be utilized to transfer to the clinic an information processing service witha considerably higher capacity and lower cost. More research is required to understand how toapply ML in a practical and secure manner. The impact of clinical ML on actual healthcare isfrequently exaggerated in the excitement surrounding it. This might be maintained by ignoranceof the factors affecting its application. Evaluations of ML methods in health care require moreeconomic analysis to make informed decisions. Informing and educating medical students in thisarea is a crucial step toward better and correct ML usage, which should be accomplished.

machine learning, deep learning, medical science, health care, medicine