Using Query Performance Predictors to Reduce Spoken Queries [chapter]

Jaime Arguello, Sandeep Avula, Fernando Diaz
2017 Lecture Notes in Computer Science  
The goal of query performance prediction is to estimate a query's retrieval effectiveness without user feedback. Past research has investigated the usefulness of query performance predictors for the task of reducing verbose textual queries. The basic idea is to automatically find a shortened version of the original query that yields a better retrieval. To date, such techniques have been applied to TREC topic descriptions (as surrogates for verbose queries) and to long textual queries issued to
more » ... web search engine. In this paper, we build upon an existing query reduction approach that was applied to TREC topic descriptions and evaluate its generalizability to the new task of reducing spoken query transcriptions. Our results show that we are able to outperform the original spoken query by a small, but significant margin. Furthermore, we show that the terms that are omitted from better-performing sub-queries include extraneous terms not central to the query topic, disfluencies, and speech recognition errors.
doi:10.1007/978-3-319-56608-5_3 fatcat:tiwpjztv3zbubms4dhbzf3gqxy