ALGORITHM FOR CLUSTERING DIFFERENT TYPES OF DRUGS AFFECTING BLOOD PRESSURE

Abstract

This article presents the development of an algorithm and software for grouping different types of drugs that affect blood pressure in humans. The results are experimentally tested on a cross-section of more than 1,100 drugs that affect human blood pressure. Using the proposed algorithm, the training sample is formed and the problem of clustering is solved. The training set consists of ten classes. Selection symbols are given in different types, they consist of nominal and value symbols. The developed algorithm works taking into account both types of symbols. The classification of learning objects is experimentally tested using a neural network-based algorithm and software.