Clustering Analysis In Data Mining Research And Application Of The Algorithm

With the rapid development of data collection and storage technologies, making thevarious organizations in the world can store vast amounts of data.data. It is because ofthe huge amount of data, making the traditional manual analysis encounter developmentbottleneck, how from massive data efficiently find meaningful information data mininghas become a major research direction and drive the rapid development of an importantfactor. As a new subject, clustering analysis technology in the data mining process in theunique position to the reality of life in the broad application, so that it has become avery active research direction.From the current situation, all the clustering algorithmsare almost always designed for a particular data object, no algorithm can do a " beapplicable everywhere", each clustering algorithm itself the advantages of andlimitations.This paper introduces data mining knowledge. After introduced the cluster analysismethod of data mining, and the cluster analysis algorithm based on the classification,detailed introduced each kind of clustering algorithm in the typical clustering algorithms,and presents the advantages and disadvantages of various typical algorithm introduced.We do a number of numerical experiments about DBSCAN algorithm, hierarchicalclustering algorithm fuzzy clustering analysis of the transitive closure algorithm and thepractical application of the combination made, and obtained some valuableexperimental results, this paper presents an improved algorithm about the traditionalK-means algorithm, and then improve the algorithm of K-means cluster stability, at thesame time, this paper proposed an improved subspace clustering algorithm, theclustering algorithm and the factors that should be considered for a suggestion.