Research On The Hierarchy Visualization Of Incremental Text Clustering

According to the characteristics of hierarchical data, visualization should meet the demand for hierarchy upstream and downstream levels and overcome the issues of visual confusion and unclear structure caused by the inherent characteristics of the hierarchy structure as well. At the same time, clustering results of incremental text clustering method change while the process of clustering which leads to reveal the character of increment in the process of visualization.On account of the demand for downstream display of hierarchical data, this paper proposes Venn diagram visualization method designed for incremental text clustering. Combined with the hierarchical structure characteristics of incremental text clustering results and the deficiency of traditionally hierarchical information visualization methods based on Venn diagram, this method sorts nodes according to the heat attribute of text and forms a sequence of nodes. Against to the no tropism of primary methods, this method also introduces the concept of direction angle to improve the array of nodes in traditional methods and visualize incremental text clustering results. Then, this method put forward incremental node arrangement which arrays new nodes to reduce changes of visualization view before and after increment and show the dynamics of incremental text clustering dynamically. Algorithm validation and demonstrate the effectiveness of this method which can array and display the process and meet the needs of hierarchy downstream and reveal the character of increment in the process of visualization.In respect of the requirements of upstream display for hierarchical data, this paper presents the force guided visualization method for incremental text clustering. Combined with the process and outcome characteristics of incremental text clustering, this method dynamically changes the force coefficient between nodes according to the deficiency of traditional force guided method, the parent nodes' level and the attribute gaps of parent nodes. This method also makes it possible to artificially change the force coefficient between current parent nodes and leaf node. After that, this method presents incremental node arrangement according to the incremental character of incremental text clustering, by which to array new nodes effectively and show the incremental character of incremental text clustering efficiently. Algorithm validation and case studies demonstrate the effectiveness of this method and highlight the differences between sub-trees and to meet the needs of hierarchy upstream.