Evaluation of Machine Learning Algorithm Performance for Medical Diagnosis Tasks

Currently, the capability of artificial intelligence algorithms to perform deep data analysis and draw conclusions from their results is being widely applied in the medical field to address various problems. One such issue is the prediction and prevention of stroke. This research study is dedicated to exploring the capabilities of artificial intelligence in detecting and predicting stroke. The research process begins with the selection of a dataset. The quality and composition of the data chosen for the study are then analyzed. Based on the analysis results, incomplete or inaccurate data are corrected and processed. After the dataset is prepared, machine learning algorithms are applied to predict stroke based on this dataset. In this study, Logistic Regression, Support Vector Machines, Random Forest, and K-Nearest Neighbors machine learning algorithms are utilized. During the course of the research, to improve the accuracy of the selected algorithms, the hyperparameters of the artificial intelligence model are fine-tuned and the model is retrained. During the training process, each algorithm is tested, and the test results are evaluated using accuracy and error indicators such as precision, recall, and F1-score. The results obtained for each algorithm are analyzed, and at the conclusion of the study, the most effective algorithms for predicting stroke are identified. The significance of this research lies in the fact that the findings obtained from the study will be applied in the medical field to enhance the possibilities of stroke prevention and early diagnosis.