THE USE OF ENSEMBLE METHODS FOR MEDICAL DIAGNOSIS PROBLEMS

The article examines the problem of developing a medical diagnosis system based on ensemble methods and evaluating its effectiveness. The system automatically performs diagnosis using disease symptoms, laboratory test results, and other patient-related data, and this capability is confirmed by results obtained on a number of datasets. It is shown that the accuracy values achieved using the ensemble method are almost equivalent to those of other algorithms, and in certain cases demonstrate superior performance. Additionally, to verify the correct operation of the software package developed using the ensemble method, the Heart Disease dataset, the Diabetes dataset, and the Multiclass Diabetes dataset were used, and their accuracy levels and corresponding confusion matrices were calculated.