Research and Improvement of TextRank Algorithm Adding Degree Adverbs

Abstract

Abstract Textrank algorithm is an approach to exact keywords in articles which will be the premise to find useful information in massive network data. After analysing the characteristics of this algorithm, this paper improves the accuracy of keyword extraction by adding degree adverbs. Based on the function of the original algorithm, a degree adverb table is established, which contains all the degree adverbs that need to be marked. By adding the degree adverbs marker, the results of keyword extraction for the same text will be improved. According to experimental studies, for most articles, adding degree adverbs judgment function increases the importance of key words in these articles. At the end of this paper, the keyword extraction results for different types of articles is demonstrated and compared.