Subspace clustering visual analysis method based on dimension correlation

Abstract

The invention discloses a subspace clustering visual analysis method based on dimension correlation. The subspace clustering visual analysis method comprises the steps of establishing a dimension correlation measurement method based on clustering significance; establishing an effective visualization method for a subspace clustering complex structure; and establishing a visual analysis framework based on the dimension correlation. In the interactive and visualized data exploration process, the invention give a user effective guidance information to guide the user to quickly find the valuable subspace and the corresponding cluster.

Description

technical field [0001] The invention belongs to the technical field of data mining and visual analysis, and relates to a visual analysis method of subspace clustering based on dimension correlation. Background technique [0002] Cluster analysis is one of the key technologies in the field of data mining. Subspace clustering is an extension of traditional clustering methods in high-dimensional data spaces, and its idea is to localize the search in relevant dimensions. [0003] Traditional clustering methods mainly encounter two problems when performing clustering in high-dimensional data sets. 1. There are a large number of irrelevant attributes in high-dimensional data sets, so that the possibility of clustering in all dimensions is almost zero; 2. The data distribution in high-dimensional space and low-dimensional space should be sparse, and the distance between data is almost equal. It is a common phenomenon, and the traditional clustering method is clustering based on d...

AI Technical Summary

Problems solved by technology

However, these methods are limited to visualization of the results of automated subspace clustering methods, and users cannot perform interactive visual analysis

[0006] The method closest to the method of the present invention is "Dimension projection matrix / tree:Interactive subspace visual exploration and analysis of highdimensional data.IEEE TVCG,19(12):2625-2633,2013" proposed by Yuan Xiaoru et al. The distance is based on the Pearson correlation coefficient, which can only be used for feature selection and lacks the meaning of clustering

In addition, for the complex structure of subspace clustering, such as the overlap of different subspaces in dimensions, the overlap of different clusters in data, etc., none of the above visualization methods give a good solution

Images