Controlling uncertainty in discretization of continuous data
Denis V. Popel, E.I. Popel · 2004
There is a demand for formalization of dependencies between multiple-valued and continuous signals. The composition of both has been successfully applied to various problems of circuit design and knowledge discovery. This paper reviews some information measures in the multiple-valued and continuous domains considering different representation forms for functions with multiple-valued and continuous attributes. Being at the cross-roads of analog and discrete data processing, the paper addresses the problem known as discretization and introduces an adaptive method of finding an optimal representation of continuous data in the multiple-valued domain. Both multiple-valued and interval decision diagrams are examined for efficient representation of continuous data.