A Clustering Algorithm Based on the Text Feature Matrix of Domain-Ontology
Guangming Gong, Jiang Yanhui, Wang Wei, Zhou Shuang-wen · 2013
The text feature matrix of domain-ontology has the following three characteristics: high-dimension, sparse and independence of dimensions. Independence means that text implications of dimensions are different from each other. Many clustering algorithms take into account the characteristics of high-dimension and sparse, but ignore the impact of independence. And the artificial interference in parameters can often affect our clustering results. In this paper, we propose a new clustering algorithm by enriching connotation of similarity and minimizing the influence of subjective parameters. The experimental results verify the validity of our algorithm.