How to find valuable references? Application of text mining in abstract clustering

The emergence of text mining has enabled firms to understand the requirement of consumers in real-time by mining publicly available information from Internet. More and more researchers pay more attention to improving the methods of text mining and the role of text mining has become increasingly prominent. This paper introduces the use of text mining in article screening and classification to help graduate students select more valuable articles and quickly have a clear understand of research status and content. Using data from a popular Chinese database CKNI, we find that thousands of articles appear when we use keywords to search in the database and these articles are complex and disorganized. Our study provide a method to screen and filter these articles, and extract topics for each class. Our study has a great significance to the improvement of graduate students' research ability in a short time.