Text Clustering Algorithm Based on Semantic Graph Structure

As semanticinformation is often missing in text representation, this paper proposes semanticgraph structure to represent text and optimize graph structure by semantic similaritymatrix. Then calculate the similarity of semanticgraph structure by using the maximum common sub-graph of graph theory. Finally, K-means algorithm will be applied to expand Chinese text clustering to improve text clustering effect. Experimental results show that the proposed algorithm based on semanticgraph structure is more conducive for the representation of semanticinformation. It significantly improves the accuracy of text similarity in the text similarity measure and will be applied to text clustering to further improve the effect of text clustering by adjusting the corresponding parameters.