Probabilistic Query Expansion method using recommended past user queries

A plenty of Query Expansion techniques have been proposed to solve the problems of information retrieval systems, but new challenges has been introduced for these methods of expansion of user queries because of the rapid growth of the size of the Web collection. In this paper we have focused our attention on the improvement of the precision of the user query by a Probabilistic Query Expansion Method. Our approach consists on the use of information contained in query logs, which includes past user queries and their clicked documents, for providing high-level recommendations. The experiment results shows that the precision of the short queries increases even if we add a few terms, but the same number of terms added don't affect the long queries precision as much as for the short queries.