Ranking Aggregation for Meta-Search Engine
Ranking aggregation is an important approach to combine information and reach an agreement between various opinions. There are many applications, such as election to select a winner from a pool of candidates based on voters’ preference profile, or produce a full ranking or preference rate over a set of web pages or online movies from users’ visiting log or historical ratings. The report proposed two voting rules – pairwise margin voting rule and probabilistic propagation-based voting rule. Both methods can capture the absolute and relative position information. Also, we present some evaluation results of the two methods to illustrate some good properties. We implemented a meta search engine, where the proposed pairwise margin is employed to aggregate the searching results from some individual search engines.
