Abstract:

Authors propose an approach for a data mining problem of identification of a characteristics subset which can separate the positive and negative observations for a classification problemwith high accuracy. The pilot studies of approach were conducted on a problem of forecasting of complications of a myocardial infarction.