SYSTEM AND METHOD FOR WORD CLASSIFICATION BASED ON HYPERLINK OF WEB TEXT

摘要

The present invention relates to a system which uses the hyperlink information of a web document consisting of hypertext to classify the type of natural language words in the web document based on the hyperlink classification. Hyperlinks are reference links representing videos, URLs, text, tags, music, pictures, programs, and files for specific words and are implemented in markup languages of web documents such as HTML. Such information enables classifying the type of the object names for words configuring hyperlinks without using a high-level natural language processing model. Utilizing the metadata information of hyperlinks enables a large amount of information to be used more easily than by an analysis using natural language processing due to the properties of the web data built by the collaboration of users.