Artificial intelligence tools are seldom used for information retrieval since classical approaches have addressed this problem in an efficient way. For large scale information retrieval, the situation is different and may necessitate more powerful methodologies. In this paper we show that indeed for large scale collections, heuristic search techniques outperform the conventional approaches in addressing information retrieval. For the purpose of supporting this statement, we have designed and implemented a genetic algorithm then a hybrid genetic algorithm for information retrieval. The effectiveness of both designed algorithms is compared to a classical method by performing empirical tests on Smart collections and random benchmarks. It appears that both the designed algorithms outperform the classical approach for large data sets and the hybrid genetic algorithm yields the best performance in terms of solution quality and runtime.
[1]
Bahgat A. Abdel Latef,et al.
Using Genetic Algorithm to Improve Information Retrieval Systems
,
2008
.
[2]
Gerard Salton,et al.
Term-Weighting Approaches in Automatic Text Retrieval
,
1988,
Inf. Process. Manag..
[3]
Hinrich Schütze,et al.
Introduction to information retrieval
,
2008
.
[4]
Dana Vrajitoru,et al.
Crossover Improvement for the Genetic Algorithm in Information Retrieval
,
1998,
Information Processing & Management.
[5]
Hsinchun Chen.
Machine learning for information retrieval: neural networks, symbolic learning, and genetic algorithms
,
1995
.
[6]
Weiguo Fan,et al.
Effective information retrieval using genetic algorithms based matching functions adaptation
,
2000,
Proceedings of the 33rd Annual Hawaii International Conference on System Sciences.