How to find and discover useful information from Internet is a real challenge in information retrieval (IR) and search engines (SE). In this paper, we propose and construct Path Trust Knowledge Graph PTKG model for assigning priority values to the unvisited web pages. For a given user specific topic t, its PTKG contains five parts: (1) The context graph G(t) = (V, E), where V is the crawled history web page set and E includes the hyper link set among the history web pages; (2) Retrieving knowledge implied in the paths among these web pages and finding their lengths; (3) Building the trust degrees among the web pages; (4) Constructing topic specific language model and general language model by using the trust degrees; (5) Assigning the priority values of web pages for ranking them. Finally, we perform an experimental comparison among our proposed PTKG approach with the classic LCG and RCG. As a result, our method outperforms LCG and RCG.
展开▼