A Hierarchical Text Clustering Algorithm with Cognitive Situation Dimensions

Text clustering is an important task of text mining. The purpose of text clustering is grouping similar text documents together efficiently to meet human interests in information searching and understanding. The procedure of clustering should involve a cognitive process of text understanding or comprehension.This paper introduces an innovative research effort, CogHTC, a hierarchical text clustering algorithm, inspired by cognitive situation models. CogHTC extracts representative features from four elaborately selected cognitive situation dimensions with consideration of the clustering efficiency. The experimental results testified good performance of CogHTC, and revealed that the clustering results of CogHTC are class or domain sensitive, and CogHTC performed better on cross-class clustering than inner- class clustering.