About a New Approach of Selection and Reduction on System Feature in Pattern Recognition
Zeng Huing-li · 1999
Based on the theoretical aspects of rough sets, the issue of the lower approximations andthe upper approximations of classification are discussed for knowledge representation of patternrecognition. The classification ability of a system feature is evaluated with the classification affectionintroduced by reducing system features and comparing the final classification result in patternrecognition. A new method of reduction and optimization on system feature is proposed. The newapproach is very valid far reducing the dimensions of knowledge representation system and dataintelligence processing.