Algorithms for Clustering of a Heterogeneous Data on the Example of Solution of the Medical Task

The paper describes the original algorithm of a heterogeneous data clustering is based on complex application of a set of measures of distances and clustering methods and multi-stage clustering. In the algorithm we use ranging of attributes the object on their importance for group and a choice of an optimum attributes set, ensemble approach to get the final clustering solution. The algorithm is realized in MixDC (Mixed Data Clustering) software system. The technique and results of the solution of a real problem of a medical data clustering in software system are described.