AN ALGORITHM FOR SELECTING AN INFORMATIVE SYMBOLIC COMPLEX BASED ON CLASSIFICATION ERROR COEFFICIENTS AND PROBABILISTIC INDICATORS IN THE REPRESENTATION OF SYMBOLS

Аннотация

In the primary processing of information, in particular in character recognition, an important issue is the selection and classification of an informative feature or a set of features classifying objects. Despite the fact that a number of methods and algorithms have been proposed to solve these problems, there are many problems in this direction that are waiting to be solved. This is due to the fact that many of the proposed approaches strongly depend on the nature of the object of study, the number of its features, the type of perceived values of features, the size of the study sample, etc., and impose certain requirements on the above. In addition, each method or algorithm will strongly depend on whether the criterion of informative selection of features and the defining rule determining the quality of the choice made are correctly chosen. This article presents a description of the algorithm developed taking into account the above approaches to the selection of information complexes of signs, as well as recommendations on the application of this algorithm in practical matters of the medical field, i.e. in ischemic heart disease obtained as an object of study (5 classes, 507 objects, 89 signs, including X_1 class “strenuous angina”, X_2 class “Acute myocardial infarction”, class X_3 “Arrhythmic form”, class X_4 “Postinfarction cardiosclerosis”, for class X_5 “Persistent form of atrial fibrillation”) formulated training was applied to the selection and positive results were achieved.