Study Of Clustering Algorithm Based On Attribut In Data Minging

Nowadays, data increase rapidly everyday. The first reason of storing such a mount of data is that the technology of computer makes it convenient, the second reason is that these data have tremendous latency function. Derived from this the concept of data mining is widely recognized, and many interrelated technologies and products are appeared. Authoritative Gartner research group reported that data mining is one of ten new technologies that are invested focally in the seven seas. This thesis starts with the concept of data mining, looks on research object of data mining in the perspective of data structure, discusses "cluster" (the important tool of data mining) in depth, and classifies cluster into Q cluster based on data element and R cluster based on attribute. The thesis emphasizes R cluster that is less involved in all kinds of literatures but has important application significance, discusses interrelated concepts . technology and algorithm. At last this thesis introduced a practical application system-the evaluate system of doctor medical treatment quality.The main aspects of this thesis are as follows:1. The cluster based on attribute effectively holds out the realization of the important capability of data mining.2 . The theory foundation about fuzzy set theory in cluster analysis application.3. The thesis discusses the fuzzy clustering algorithm based on attribut.4. The thesis discusses the future work aspects and facing challenge of this task .In general, this thesis is based on the basic theory and technology of cluster analysis in data mining tool, puts forward some new viewpoint and algorithm designing notation and attempts to discuss in theory and practice, combining the present practical application.