Determination of predisposition to diabetes mellitus based on Fuzzy neural network

Abstract

This article solves the problem of determining the occurrence of diabetes mellitus in pregnant women. The aim of the study is to find an effective method based on the machine learning method, which will simplify the procedure and increase the speed of determining this disease with a high degree of accuracy. The currently existing methods for determining diabetes mellitus cannot allow constantly monitoring the presence of diabetes mellitus in pregnant women due to the complexity of the procedure, as well as the long time it takes to process the analysis results. Therefore, the use of modern technologies, in particular machine learning methods, will get rid of these disadvantages and allow to continuously monitor the possible occurrence of the disease. In this article, the subject area is analyzed and the relevance of the research topic is considered, the initial data preparation and collection for a neuro-fuzzy model constructing are carried out. In addition, intelligent models based on various machine learning methods have been constructed. The best method based on a fuzzy neural network has been chosen, which allows to classify the available data with a high degree of accuracy. A use-case diagram for solving practical problems of determining the presence of diabetes mellitus has been developed.