A Multi-document Summarization Method Based on Frequent Itemset

摘要

Abstract-Multi-doeument summarization can convey the information contained in many articles with the same topic. This paper proposes a multi-document summarization method based on frequent itemsets. Firstly, this method takes multi-document as a transaction set, and frequent itemset mining is adopted. Then it classifies the multi-document sentences according to the frequent itemsets, and divides the multi-document into sub-topics. Finally, we can obtain the multi-document summary from combing the summary sentences extracted from the sorted subtopics. In the experiment, we select 20 multidocument data sets, the summaries generated by this method has more comprehensive and higher coverage of the original multi-document than that of MEAD.