<![CDATA[Strokes are a significant health problem because they often lead to long-term disabilities due to delayed diagnoses and insufficient information about the disease. The use of artificial intelligence (AI), specifically machine learning (ML) and deep learning (DL), has the potential to aid in stroke diagnosis and significantly advance healthcare. This review article critically examines predictive methods for ischemic and hemorrhagic strokes. The preferred reporting items for systematic reviews and meta-analyses (PRISMA) method was used to identify 79 relevant articles from five databases spanning 2012 to 2022, with IEEE having the highest number of articles and citations. China had the most authors, and the random forest (RF) algorithm showed the most accurate results. A taxonomy categorizing the implementation and usage of ML and DL for stroke prediction was created and includes five focus areas: building, system planning, evaluation, comparison, and analysis. Additional research into other disease features related to stroke is warranted. Decentralized federated learning should also be implemented to collect data from remote locations for early diagnosis and create a single training model.]]>
Copyrights © 2024
