Combining Relevance Clustering and Graph Model Methods for Automatic Summarization

摘要

With the exponential growth of Internet information, the application value of automatic summarization continues to improve. Automatic summarization is an important research direction, which based on the relationship between sentences. This paper proposes a new automatic summarization method combining clustering and graph model, details adding the cluster center features generated by density peak algorithm into a graph model summarization system. The method aims to apply the thought of distance constraint extracting from the cluster center to Chinese automatic summarization, so as to reduce redundant. Through the evaluation on the corpus of scientific and technological achievements documents, it shows that the system to a certain extent, improves the summarization quality of the document.