Abstract: Automatic abstracting has been a priority research point in computational linguistics field, and the study and application of automatic summarization have widely attracted the attention of interrelated academic subjects such as computer science, linguistics, informatics. I}his article firstly brought out how LexRank algorithm works in automatic summarization, then improved the method in three aspects including sentence similarity computing, sentence weight computing and redundancy resolution. And the factors of influence could be dynamically adjusted according to the documents content. The system described in this article could deal with single or multi-document summarization both in English and Chinese. With evaluations on two corpuses, our methods could produce better summaries than the original LexRank algorithm to a certain degree. We also show that our system is quite insensitive to the noise in the data that may result from an imperfect topical clustering of documents. And in the end, existing problem and the developing trend of automatic summarization technology were discussed.

Key words: Automatic abstracting, LexRank, Sentence similarity, Dynamic adjustment, Redundancy resolution