Research On Clustering Algorithm Based On Co-evolutionary

Data mining is the process of extracting people interested in knowledgeand information from a large, incomplete, fuzzy and random noise data. Theinformation and knowledge is implied, which is not been known beforehand,potentially useful information. Clustering analysis is one of the mostpromising methods. After years of development, clustering analysis issolid theoretical foundation and forms the system methodology system.However the traditional clustering analysis method is confined to thetheoretical analyses and depends on the probability distributioncharacteristics of the data，its actual application less considering datacharacteristics and differences。Therefore, how to overcome traditionalclustering analysis method for the probability of hypothesis, the academicresearch for that become hot in recent years.Aiming at above problems, this paper investigates the cooperativeevolutionary clustering analysis method. Based on the mechanism ofordinary evolutionary algorithm analysis and ordinary cluster analysismethod, we design co-evolution and co-evolutionary algorithms mechanism.Based on the cooperative co-evolution ideas, this paper present a kind ofco-evolution clustering analysis method. A number of experiments verifythe validity of this method.The end of this article, one uses co-evolutionary clustering algorithmto hierarchies division and rating for the original scores of students’grades. The traditional rating method is based on the evaluation andabsolute scores, and this method has some defects. According to thetraditional method to evaluate the students’ grades will be biased, butalso not is reasonable and effective evaluation of teachers’ teachingeffect. Therefore this paper reference data mining in the cluster analysis,based on cooperative evolutionary to use the cluster analysis algorithmfor students of original scores rank, can effectively overcome the defectof traditional evaluation method.