Abstract: The paper addresses issues of special style structuring of learning set in pattern recognition area. Above the regular means of ranking of objects and properties, which also use the structure of learning set, the logic separation hypotheses is treated over the multi value features area, which structures the learning set and tries to recover more valuable relations for better recognition. Algorithmically the model is equivalent to constructing the reduced disjunctive normal form of Boolean functions. The multi valued case which is considered is as harder as the binary case but it uses approximately the same structures.

Keywords: Learning, Boolean function, logic separation.

ACM Classification Keywords: F.2.2 Nonnumerical Algorithms and Problems: Computations on discrete structures

Link:

ON STRUCTURAL RECOGNITION WITH LOGIC AND DISCRETE ANALYSIS

Levon Aslanyan, Hasmik Sahakyan

http://foibg.com/ijita/vol17/ijita17-1-p01.pdf