Improving Query Expansion with Stemming Terms: A New Genetic Algorithm Approach.
Lourdes Araujo, Jose R. Pérez Agüera
Proc. European Conf. on Evolutionary Computation in Combinatorial Optimisation (EvoCop 2008).
LNCS 4972, pp. 182-193, Springer-Verlag (2008).

Nowadays, searching information in the web or in any kind of document collection has
become one of the most frequent activities. However, user queries can be formulated
in a way that hinder the recovery of the requested information. The objective of automatic
query transformation is to improve the quality of the recovered information. This paper
describes a new genetic algorithm used to change the set of terms that compose a user
query without user supervision, by complementing an expansion process based on the use
of a morphological thesaurus. We apply a stemming process to obtain the stem of a word,
for which the thesaurus provides its different forms. The set of candidate query terms
is constructed by expanding each term in the original query with the terms morphologically
related. The genetic algorithm is in charge of selecting the terms of the final query
from the candidate term set. The selection process is based on the retrieval results
obtained when searching with different combination of candidate terms. We have obtained
encouraging results, improving the performance of a standard set of tests.