The Research And Implementation Of Automatic Summarization Algorithm Based On Sentence Weight And Chapter Structure

Automatic summarization is to get the central content or the valuable component for customers from original document by processing text information on computer. Automatic summarization offers people a simpler and more efficient way for obtaining information.This dissertation proposed a kind of automatic summarization algorithm based on sentence weights and chapter structures. First of all, with the sentence information statistical algorithm, we can grasp accurate statistics of sentences information in the article based on the interception character clause algorithm with cursor. Then, we have refined and improved the word’s score formula and the sentence’s weight formula which can make a more valuable sort of the sentences on the basis of precise statistics information of words and sentences and the segmentation of article structures respectively. Finally, when extracting the summarization sentences, we introduced the Chinese sentence similarity calculation algorithm which is based on the layers to make the smaller redundancy and correlation of chosen sentences. In addition, when choosing the summarization sentences, the users can set up the preferences phrases according to their own interests to get the more targeted automatic summarization.The experimental results show that the automatic summarizations from the algorithm proposed in this paper are better than those from most of the same type algorithms, as well as those from the Microsoft Word 2003’s summarization tool in the precision rate and recall rate. It proves that the algorithm can provide a good automatic summarization for users.