Automatic Query expansion based on tag recommendation

Abstract

Query expansion is the process of adding new terms to queries posed by users in an attempt to improve search results. This is especially useful when the original query is not able to unequivocally express the user information need. Traditional query expansion approaches are based on pseudo-relevance feedback. They select expansion terms from highly ranked documents retrieved as result of the original query, assuming that these documents contain useful terms for query expansion. However, this might not always be the case, especially for queries considered 'difficult'. Recently, some query expansion techniques exploit external sources of information found on the Web, such as on-line encyclopedias and collaborative social annotation systems, to select semantically rich expansion terms. We here propose a new method for expanding queries related to entities (or narrow concepts). More specifically, we propose a method for automatically filtering, weighting and ranking terms extracted from Wikipedia articles related to the original query. Our method is based on state-of-the-art tag recommendation methods that exploit heuristic metrics to estimate the descriptive capacity of a given term. Originally proposed for the context of tags, we here apply these recommendation methods to weight and rank terms extracted from multiple fields of Wikipedia articles according to their relevance for the article. We evaluate our method comparing it against three state-of-the-art baselines in three collections. Our results indicate that our proposed method outperforms all baselines in all collections, with relative gains in MAP of up to 15% over the best baselines.