The Chinese Vietnamese bilingual news event ranking method based on attribute association graph

Facing with bilingual news event's attributes association about Chinese and Vietnamese, we propose a bilingual news event sequencing method based on graph of attributes association. This method built the graph of attributes association based on the relations between event properties and it sequences the Chinese and Vietnamese bilingual news event using this graph. This method first extracts the titles, elements, topic sentences which characterize bilingual news event properties. And it uses the extractions to describe events in the document. Then we translate these title, elements and topic sentences by using the bilingual dictionary and Vietnamese-Chinese Alignment Corpora. And we build the graph of attributes association. We calculate the weights of edges by using word2vec and elements co-occurrence strength. In the end, it sequences the bilingual news using similarity among event node, query keyword and connection between the graph's nodes and event. The experimental results show that the proposed method can effectively improve the performance of the ranking of the news events in the Chinese and Vietnamese, and the attribute relevance has a good effect on the ranking of the Chinese and Vietnamese news events.