Community structure of the Chinese document network based on content similarity

Based on the complex network theory, we proposed a clustering algorithm based on content similarity. Firstly, the Chinese documents are represented by the vector-space model, and the content similarity between any two documents is computed by the cosine similarity. Consequently, the network node is defined as a document, and the edge weight is defined as the similarity obtained by the cosine similarity definition. The document connectivity network can be constructed based on the document-to-document similarity graph. If the edge weight between any two nodes is smaller than a constant value, then it´s set as zero. Using the edge betweenness of the network, we reconstructed the hierarchical structure of the funding proposal network. Computing the edge betweenness, and remove the edge with largest betweenness; Repeat the above process until all edges are removed. Using an open dataset proposed by Fudan University, we experimentally compared the performance of the partition clustering algorithm and other algorithms, such as K-means and Bisecting K-means. The numerical results indicate that our algorithm is more efficient than K-means and Bisecting K-means algorithms. In addition, the numerical results are robustness to different constant. Finally, the algorithm is implemented on the proposal network, the community structure based on the content similarity is detected.