Single Chinese News Article Summarization Based on Ranking Propagation

This paper proposes a news summarization system called NewsSum for simultaneous key entities and sentences extraction from single Chinese news article. NewsSum can provide both query-independent and query-specific news summarization. In this study, NewsSum is implemented by firstly parsing the news text in the preprocessing stage and building a news document graph, then exploiting the ranking propagation algorithm on the graph to extract key entities and sentences from the text as its summary. Furthermore, the quality of NewsSum is compared with state-of-the-art summarization approaches through a user survey.