Improved algorithms for calculating evaluations in processing medical data

Abstract

The paper examines the issues of diagnosis and treatment of cardiovascular diseases, commonly encountered in diagnostic decision-making, when medical data are processed. The issues of classification of heart diseases and detection of informative signs are solved on the basis of estimation algorithms. In addition, the appropriate software was developed. The main goal of the research is to solve such issues as constructing inter-object remoteness in a complex of informative features that distinguish objects of diagnostic classes, select a complex of signs that characterize mutual differences of objects, and also identify the value of the proximity function when diagnosing an unknown object [1-5]. The level of significance or representation of the set belonging to the j-object ofХр-class, which are the main stages of the algorithms for calculating its assessment relative to the class [1-5], was revealed. An algorithm for diagnosing an unknown object in the space of informative features was proposed. The suggested theoretical ideas were confirmed in practice. In addition, the decision rules in this space and their software were developed [4-5].