Chinese Keyword Extraction Using Semantically Weighted Network

The complex network theory is widely used in the field of keyword extraction. Through analyzing the insufficient of keyword extraction algorithms using traditional complex network, this paper proposes a new method to extract Chinese keyword based on semantically weighted network. On the basis of K-nearest neighbor coupling network, we build semantically weighted network according to the co-occurrence frequency and semantic similarity of the words in the text. We calculate the betweenness value, clustering coefficient variation and shortest path variation of the word node in the network to obtain the comprehensive eigenvalue of each word. According to the size of comprehensive eigenvalue, we extract text keyword. The experimental results show that the keywords extracted by this method can reflect the theme of the text better, and the accuracy has been significantly improved.