Machine Learning Algorithms in Recommendation System for Diagnosis of Breast Cancer According to Microwave Radiothermometry

This study is devoted to the development of a recommendation system for early diagnosis of breast cancer according to microwave radiothermometryMicrowave radiometry is a modern method of medical diagnostics, based on measuring the intrinsic electromagnetic radiation of human tissues in the infrared and microwave wavelengths. Over the past two decades, this method has proven itself in a number of branches of medicine. At the same time, a number of problems were identified. One of the most relevant of them is the development of methods and algorithms for the qualitative analysis and interpretation of thermometric data. In this paper, we propose a solution to this problem based on the creation of an appropriate recommendation system. First of all, it provides an effective solution to the classification problem, i.e. assignment of the examined patient to the corresponding diagnostic class. At the same time, the recommendation system being developed provides a qualitative interpretation of the obtained thermometric data and the justification of the proposed diagnostic solution in the user's language, i.e. diagnostician.A distinctive feature of the classifiers under development is a specific feature space built on the basis of mathematical functions simulating the nuances of the behavior of the temperature fields of patients of various diagnostic classes.