In this paper, we describe the methodology and experimental analysis of a twofold strategy for the retrieval of medical relevant information: a ranking fusion and a query reformulation approach. In particular, the query reformulation approach is based on the idea that a query is composed of two parts: the primary term and the secondary term of the query, and that these two parts can be substituted with alternative terms to create a reformulation of the original query. The goal of our work is to evaluate the performances of a search engine over 1) manual query variants; 2) different retrieval functions; 3) w/out pseudo-relevance feedback; 4) reciprocal ranking fusion. We describe the experiments based on the CLEF eHealth 2021 Consumer Health Search Task dataset. The results show that 1) a ranking fusion approach of the baseline models improves MAP significantly; 2) manual query variants open new questions about possible an unintentional bias in the pool of documents that were selected for relevance assessment.
Did I Miss Anything? A Study on Ranking Fusion and Manual Query Rewriting in Consumer Health Search
Di Nunzio G. M.;Vezzani F.
2022
Abstract
In this paper, we describe the methodology and experimental analysis of a twofold strategy for the retrieval of medical relevant information: a ranking fusion and a query reformulation approach. In particular, the query reformulation approach is based on the idea that a query is composed of two parts: the primary term and the secondary term of the query, and that these two parts can be substituted with alternative terms to create a reformulation of the original query. The goal of our work is to evaluate the performances of a search engine over 1) manual query variants; 2) different retrieval functions; 3) w/out pseudo-relevance feedback; 4) reciprocal ranking fusion. We describe the experiments based on the CLEF eHealth 2021 Consumer Health Search Task dataset. The results show that 1) a ranking fusion approach of the baseline models improves MAP significantly; 2) manual query variants open new questions about possible an unintentional bias in the pool of documents that were selected for relevance assessment.File | Dimensione | Formato | |
---|---|---|---|
dinunzio_vezzani_clef2022.pdf
Accesso riservato
Tipologia:
Published (publisher's version)
Licenza:
Accesso privato - non pubblico
Dimensione
1.34 MB
Formato
Adobe PDF
|
1.34 MB | Adobe PDF | Visualizza/Apri Richiedi una copia |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.