Automated quality assessment of web pages from textual content

Given the vastness of Internet, search engines have to find not only relevant but also high-quality web pages to satisfy users´ information need. At present, most quality assessing methods for web pages are based on link analysis and user feedbacks. Considering that users acquire information from web pages mainly through reading their text, this paper addresses automated quality assessment of web pages from textual content. This paper surveys related works on assessing text´s quality, summarizes quality-related features, and examines them with a real-word data set. Experimental results show that features based on the length of text are the most effective, while combining length features with other features such as part-of-speech tags and readability can further improve the accuracy.