Unsupervised Classification Approach to Developing a Medical Diagnosis Based on the Results of Prepared Tests

A formal procedure is proposed for making a medical diagnosis based on the results of analyzes that satisfy the specified requirements: for each type of possible disease, a set of analizes is formed, and in each such set the analyzes are ranked by "importance" for this disease. Presentation is conducted on the example of a specific sample of Dermatology from the known repository of UCI Machine Learning. The results indicate that the proposed procedure may be a useful medical assistant in making a definitive diagnosis.