Sustainability of dividing data in intervals in the problems of recognition and searching for hidden laws

A great role for the improvement of digital technologies in the scientific, industrial and social spheres has the search for new knowledge contained in databases and data warehouses in the form of hidden patterns. In this paper, in order to identify hidden patterns in the recognition of objects, the method of splitting the characteristic values into disjoint intervals is considered. As a quality criterion, it is proposed to use the value of the stability indicator for dividing the original and latent features into intervals.