MOGP Strategies for Topical Search Using Wikipedia

Abstract

Genetic Programming techniques have demonstrated great potential in dealing with the problem of query generation. In order to assist the user with thematic recommendations, this work explores different Multi-Objective Genetic Programming strategies for evolving a collection of topical Boolean queries. This study compares three approaches to build topical Boolean queries: using terms, incorporating Wikipedia semantics (Wikipedia concepts) and a hybrid approach, using a combination of both terms and concepts. In addition, different fitness functions are combined giving rise to seven multi-objective schemes. In particular, we propose novel fitness functions aimed at attaining high diversity based on the information-theoretic notion of entropy and Jaccard similarity. Experiments were completed using 25 topics from a dataset consisting of approximately 350,000 webpages classified into 448 topics. The results reveal that there are no statistically significant improvements in efficiency when terms, concepts or a combination h of both is used. However, the use of terms allows to discover rartificial queries that are hard to interpret by the humans. On the contrary, the use of concepts have a positive effect on interpretability and simplicity (considering the number of operands), resulting in better execution times. In ddition, p several differences are observed when using different combinations of fitness o functions.