Rank aggregation algorithm using particle swarm optimization for metasearch engines

Rank aggregation is the algorithm that generates a consensus ranking from a given set of rankings from different sources. Rank aggregation algorithm has been found in many applications in web including metasearch and spam fighting. The proposed rank aggregation algorithm is called footrule optimal aggregation (FOA) because it is based upon optimizing the spearman footrule distance using particle swarm optimization. The proposed rank aggregation algorithm is implemented as a part of a complete metasearch engine. In order to ensure the effectiveness of our proposed rank aggregation algorithm, we compare the results obtained by our proposed PSO based algorithm with that obtained by the classical Borda's count method. The obtained result shows that our proposed rank aggregation method is better than the classical Borda's based method in two sides. The first side, our proposed PSO based algorithm gives lower footrule distance value than results obtained by Borda's based method for the same query .The second side, our method satisfy Condorcet criteria whereas, the Borda's based algorithm does not. On the other hand, the Borda's based rank aggregation algorithm has relatively lower running time than our proposed PSO based one.