Web Query Reformulation Using FireFly Algorithm

Abstract

Searching for information on the Web engages the user in a process of questioning for the choice of search engines. However, many Internet users suffer for the information choice which these search engines receive. On the other hand, if the queries do not express their needs or else their objectives, this implies that some information is not formulated, requiring the reformulation of these queries. In this paper, an approach of bio-inspired optimization based on the FireFly Algorithm is used to formulate the query by providing a new suggestion. This algorithm has been applied on the frequent itemsets generated by FP-Growth (frequent-pattern Growth). Moreover, every user interaction with the search engine has been treated as a Firefly path. The algorithmic solution allows the user to select the best path (which contains key words) among all possible solutions for the initial query. Experimentally, we study the performance of the proposed method in comparison to different techniques (particle swarms optimization and genetic algorithms).