Title :
Finding reducts with user specified criteria
Author_Institution :
Dept. of Comput. Sci., Univ. of Wales, Cardiff, UK
Abstract :
Rough set theory is emerging as a powerful tool for inducing classification knowledge from databases. Central to this technique is the problem of searching for a reduct, which is a minimal subset of information that has the same ability to classify data as when the full set of information is used. In this paper, we introduce a method for finding reducts based on two user-specified criteria. This is in contrast to many existing techniques which use user-independent criteria for measuring information redundancy. Our approach allows flexibility in knowledge extraction and is capable of dealing with different levels of noise in data for different applications
Keywords :
database theory; deductive databases; knowledge acquisition; noise; pattern classification; redundancy; set theory; classification knowledge induction; data noise levels; databases; flexibility; information redundancy; knowledge discovery; knowledge extraction; minimal information subset; reduct searching; rough set theory; user-specified criteria; Active noise reduction; Computer science; Data mining; Electronic mail; Noise level; Noise measurement; Redundancy; Rough sets; Signal to noise ratio; Spatial databases;
Conference_Titel :
Database and Expert Systems Applications, 1997. Proceedings., Eighth International Workshop on
Conference_Location :
Toulouse
Print_ISBN :
0-8186-8147-0
DOI :
10.1109/DEXA.1997.617309
