Design And Implementation Of Automatic Summarization System Based On Textrank Algorithm

Nowadays,the Internet has been developing rapidly,people are more and more dependent on the network in their daily life,The Internet is gradually being the main medium for people to obtain and disseminate information resources.The rich Internet resources offer convenience for people's life,but at the same time,it also eats into lots of time of people to filter out what they need,excessive information brings a more serious problem to people.The automatic summarization technology can extract and compress the text information automatically by computer,which greatly reduces the reading cost and gradually becomes the most effective method to deal with the problems caused by the massive network information.As soon as the automatic summarization technology is put forward,it has attracted a steady stream of scholars in the world to carry out research,and its implementation method has gradually progressed and matured with the process of history,which has an immeasurable application prospect.This paper mainly through the investigation and analysis of various implementation methods in the development of automatic summarization technology,especially the in-depth study of the automatic summarization method based on the TextRank algorithm,and found that the TextRank algorithm failed to making full use of the text features in sentences when construct the graph model and calculate the sentence similarity,there is a lot of room for improvement.So in this paper,a advanced automatic summarization algorithm based on TextRank algorithm is proposed,which mainly considers more text features,such as clue words,sentence location information,text topic,etc.,rather than simply based on statistics when constructing graph models and calculating sentence similarity.Then,related experiments were designed on the public Chinese summarization dataset of NLPCC 2017 shared task3 combined with ROUGE evaluation indicators to verify the effectiveness of the automatic summarization algorithm proposed in this paper.Finally,an automatic summarization web system which use the algorithm proposed by this paper is designed and implemented,which mainly displays the process and result of automatic summarization.At the same time,the text retrieval function is realized based on the million-level text data crawled by Internet crawlers,the search results are displayed in the form of text titles and automatic summaries,reducing users' reading costs.This system is designed to allow users to experience the process and basic applications of automatic abstracting technology on the web.