Research On Application Of Rough Set Theory In Data Mining

Rough set is a good mathematical tool to deal with uncertain information. In recent years,there have been more and more applications of rough set theory in the field of data mining.Clustering and attribute reduction are always important and difficult in data miningtechnology, how to further improve the utilization of the rough set theory in this field is verymeaningful.The main research works are as follows:1. After given the basic definition of the cluster and its method，points out theshortage of the classic k-means algorithm. Taking advantage of the rough set theory inexcellent boundary data processing and combining with the good global searching ability ofPSO, puts forward a rough clustering method based on improving particle swarm algorithm.The experiment results show that the clustering method in processing boundary data ofclustering has stronger ability, compared with other methods, the accuracy and stability arehigher.2. Introduced two commonly used attribute reduction methods and briefly analyzed eachcharacteristics of representative algorithm, and then points out the insufficiency of mostattribute reduction methods based on core attribute can’t deal with incomplete informationsystem , then uses limited tolerance relation structure a modified discernibility matrix, and onthis basis puts forward the attribute reduction method based on the limited tolerancerelation , and finally through the examples which comparison with the similar methods showsthat the algorithm method can assure reduction is the relatively minimal reduction .