A clustering result semantic feature extraction and visualization method based on a strong item set

Abstract

The invention belongs to the technical field of computer information processing, and provides a clustering result semantic feature extraction and visualization method based on a strong item set. According to the method, firstly, a cluster semantic feature model based on a strong item set is constructed through analysis, the feature of each cluster can be described visually and effectively, and theinterpretability of a clustering result set is improved; Then a cluster semantic feature extraction algorithm CLCE based on a strong item set is provided, the algorithm has high performance, and cluster semantic features oriented to the strong class set can be effectively extracted; And finally, a cluster semantic feature visualization method is provided, a domain expert is further helped to understand each cluster in the clustering result set, and the application of the mined related knowledge model is promoted.

Description

technical field [0001] The invention belongs to the technical field of computer information processing, and proposes a semantic feature extraction and visualization method of a clustering result based on a strong item set. Background technique [0002] Clustering is one of the widely used technologies in the field of data analysis. It is a common method to analyze data with the idea of ​​"like flock together" without pre-specifying categories. However, the result of clustering—the interpretability of the clustering result set is the key to the success of clustering analysis methods at the application level. Therefore, for many clustering applications, the interpretation and visualization of clustering results is more important than the clustering itself. Although there are many existing clustering models, many new clustering research results have appeared in many aspects such as basic clustering algorithms, semi-supervised clustering, time series clustering, graph or networ...

AI Technical Summary

Problems solved by technology

[0003] At present, there are some patents related to the interpretation and visualization of clustering results. Patent 201010194391.7 proposes a method for visualizing clustering analysis results, which realizes the clustering results of data information objects, the structural relationship between clustering categories and Its association and the semantic similarity between data information objects can be visually expressed, but it cannot reflect the semantic characteristics of each cluster in the clustering result set; a Radviz-based fuzzy clustering result visualization method proposed in patent 201610341872.3 is mainly based on Radviz with multi-dimensional matrix The form realizes the visualization of the distribution of membership, the size of clusters, the relationship between clusters, etc., and lacks the visualization in clustering semantic information; patent 201810255690.3 proposes a feedback clustering based on cluster semantic feature analysis Class method, which mainly uses K-means clustering method, does not design clustering result interpretation and visualization technology for any clustering method

Due to the lack of modeling and analysis of the semantic feature information of the clustering results of the general clustering method, the above patents cannot effectively describe the semantic information of each cluster in the clustering result set and enhance the interpretability of the clustering result set. Difficulties in recognizing and understanding clustering result sets

Images