Applications Of K-means Weighted Clustering Ensemble Model In App Market

Cluster analysis as an important method in data mining,has been widely used in financial analysis,market analysis,image processing and other industries.As a kind of unsupervised statistical learning method,clustering in disorder or no a priori information of object processing is of great significance.Especially in the era of big data,data has a large amount of value and the characteristics of low density,clustering analysis is increasingly important.Clustering analysis development more rapidly,the algorithm is rich.Clustering fusion algorithm from A.L.F red and A.S launched the trel,due to its better than general clustering method has the characteristics of more and more get people's attention.With the rapid development of mobile Internet,mobile phone Application software(App)numerous,quality is uneven.This paper first introduces the k-means,k-means of weighted clustering and clustering convergence of some classic algorithm,based on the use of k-means the weighted clustering fusion method,an improved fusion phases clustering method.Then select on January 1,2016 to December 31,2016 App data,data transformation,data form to represent the value of the App section.Again according to the characteristics of the App data,using correlation analysis,select the contact performance App capacity variable average size,average retention and performance App brand ability variables start times daily,average visit time and so on.Finally using the traditional k-means,k-means weighted clustering and clustering fusion method and improved clustering method to classify the App data in stages,to evaluate the classification structure.Concluded that the improved phases of clustering algorithm and clustering structure more in line with the actual situation of the App.Finally,the classification of the App structure characteristics is extracted and put forward the relevant Suggestions.