Extracting Textual Information from Website Using Mixed Rules

Textual information extraction refers to extracting meaningful information from text documents. Text is one of the essential carriers of information. Mining textual information can help people obtain knowledge and make better decisions. In today’s world, web pages contain more textural information than other sources and are more easily processed by computers. At the same time, the structure of web pages is becoming increasingly complex, making it more challenging to extract meaningful textual information from web pages than before. In this paper, we first reviewed the existing methods for extracting textual information from web pages, then discussed an assumption about the classification of web page layouts, and then proposed a mixed extraction system which aims at extracting textual information from a randomly given website by first classifying pages and then adopting a different extraction strategy for each type of page. Finally, we test this extraction system on web page from different sources. The result shows that the system has a certain level of accuracy for page classification tasks. It also has a certain level of extraction effectiveness for pages that are correctly classified.