PROJECT TITLE :

Large-scale Location Prediction for Web Pages - 2017

ABSTRACT:

Location information of Web pages plays an important role in location-sensitive tasks like Web search ranking for location-sensitive queries. But, such information is typically ambiguous, incomplete, or maybe missing, that raises the problem of location prediction for Web pages. Meanwhile, Net pages are huge and often noisy, which cause challenges to the majority of existing algorithms for location prediction. In this paper, we have a tendency to propose a unique and scalable location prediction framework for Web pages based on the question-URL click graph. In specific, we have a tendency to introduce an idea of term location vectors to capture location distributions for all terms and develop an automatic approach to learn the importance of every term location vector for location prediction. Empirical results on a giant URL set demonstrate that the proposed framework significantly improves the placement prediction accuracy comparing with numerous representative baselines. We tend to more offer a principled manner to incorporate the proposed framework into the search ranking task and experimental results on a industrial search engine show that the proposed method remarkably boosts the ranking performance for location-sensitive queries.