A Website Source Evaluation Algorithm Based on Comprehensive Feature Analysis

Traditional web page sorting algorithms can only find the single web page that is the most relevant to keywords, but can not find the relevant website information source. For tackling the problem, we Propose a website information source evaluation algorithm based on comprehensive feature analysis. This algorithm first obtains multiple web pages corresponding to keywords through Baidu and other search engines, then obtains the contents of corresponding website information sources through crawler program and extracts the features, and finally obtains the sorting results of information sources of relevant websites by calculating relevancy combining BM25 algorithm and cosine distance. At the same time, combined with the implicit feedback behavior of users' browsing time, the sorting results could be dynamically adjusted to make the search results personalized. Experiment results show that this approach could make full use of web features, and improve the quality of web source evaluation algorithm by combining the semantic information of web content.