Design And Implementation Of Automatic Summarization System Based On Word2Vec

With the rapid development of information technology and Internet, the amount of information on the Internet is growing exponentially, and the update speed is getting faster and faster.How to get useful information in the mass of information becomes increasingly important. Abstract, as an overview of the contents of text information, is the main content of the information,and it can be summarized objectively so that people can get the required information efficiently through the concise and readable abstract text. Deep learning technology can provide useful basis for the automatic text summarization.Word2 Vec, with its high efficiency training word vector characteristics, has maintained many concerns by researches. This thesis proposes a novel summarization approach based on Word2 Vec, and the main works are as follows:1) This paper proposes an approach based on term characters. The term abstraction is done on the basis of the method on subject word extraction based on features. By counting the term frequency, analyzing part of speech and term location, the key terms can be done.2) This paper proposes an automatic summarization method based on Word2 Vec. The method is based on the extraction of topic words and then weight the results. Then, this paper also evaluates the candidates of sentences on the aspects of weights.3) This paper proposes an evaluation method for evaluating the summarization performance. In detail, the experimental corpus and the extracted summarization are indexed by Lucene, and then the system can retrieve the corresponding results. The performance is evaluated by analyzing the recall ratio.Experimental results show the feasible of the approach.In the end of this paper, the existing problems and future research work plan are described.