Linguistic features to predict query difficulty

Josiane Mothe (1, 2) , Ludovic Tanguy (3, 4)

1 IRIT-SIG - Systèmes d’Informations Généralisées
2 UT2J - Université Toulouse - Jean Jaurès
3 ERSS - Equipe de Recherche en Syntaxe et Sémantique
4 CLLE-ERSS - Cognition, Langues, Langage, Ergonomie

Query difficulty can be linked to a number of causes. Some of these causes can be related to the query expression itself, and can therefore be detected through a linguistic analysis of the query text. Using 16 different linguistic features, automatically computed on TREC queries, we looked for significant correlations between these features and the average recall and precision scores obtained by systems. Three of these features are shown to have a significant impact on either recall or precision scores for previous adhoc TREC campaigns. Each of these features can be viewed as a clue to a linguistically-specific characteristic, either morphological, syntactical or semantic. These results also open the way for a more enlightened use of linguistic processing in IR systems.

Mots clés: Difficult query; linguistic features; TREC

Domaines: Linguistique; Traitement du texte et du document; Recherche d'information [cs.IR]

Josiane Mothe, Ludovic Tanguy. Linguistic features to predict query difficulty: a case study on previous TREC campaigns. ACM Conference on research and Development in Information Retrieval, SIGIR, Predicting query difficulty - methods and applications workshop, 2005, Salvador de Bahia, Brazil. pp.7-10. ⟨halshs-00287692⟩