An integrated clustering method based on evidence reasoning for user behavior analysis

Abstract

The invention provides an integrated clustering method based on evidence reasoning for user behavior analysis. The method can fully consider the time characteristics of user data and the credibility of base clusters, and solve the problems of weak robustness and stability of single clusters and poor adaptability of existing integrated clustering methods by using evidential reasoning method, so asto improve the clustering effect of user behavior data. The invention has the advantages that: the problem that the traditional clustering algorithm fails due to the high dimension of the user behavior data can be overcome; the invention can solve the problems of weak robustness and stability of single clustering machines and poor adaptability of existing integrated clustering methods, so as to improve the clustering effect of user behavior data. The invention can be used for clustering of user behavior data, in particular for clustering of user behavior data with high-dimensional characteristics, and can also be used for clustering of stream data and the like, and has wide application range.

Description

technical field [0001] The invention relates to the technical field of clustering methods, in particular to an integrated clustering method based on evidence reasoning for user behavior analysis. Background technique [0002] Currently, there are five types of clustering methods commonly used, including partition-based clustering methods, hierarchical-based clustering methods, hierarchical-based clustering methods, density-based clustering methods, and grid-based clustering methods. Partition-based clustering methods, representative methods such as-means (k-means) clustering method, its idea is to divide the objects closest to the cluster center into a cluster; the idea of ​​hierarchical clustering method is through A method for clustering by creating a hierarchical decomposition for a given set of data objects; a density-based clustering method, such as the DBSCAN algorithm, which assumes that the cluster structure can be determined by the tightness of the sample distributi...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above-mentioned technical defects in the prior art, the present invention provides an integrated clustering method based on evidence reasoning for user behavior analysis, which can fully consider the time characteristics of user data and the credibility of the base clusterer, The problem of weak robustness and stability of a single clusterer and the poor adaptability of existing integrated clustering methods are comprehensively solved by adopting the method of evidential reasoning, thereby improving the clustering effect of user behavior data

Images