Do frequent media words worsen query expansion? - Archive ouverte HAL Access content directly
Conference Papers Year : 2017

Do frequent media words worsen query expansion?


This paper offers a linguistic approach to the study of the potency of query expansion while retrieving information from the web. The expansion allows enhancing the results; however, some queries show lower effectiveness after expansion. The objective of the study is to analyze linguistic features of initial query (IQ) as predictors for the expansion potency by different systems. The IQ is considered as a ‘bag of words’ with their linguistic descriptions, frequency first of all. The interdependence of different linguistic features of a query term determines the term value and its validity for the expansion. Analyzing two sets of terms from IQ (from queries that failed and from queries that were improved after expansion), we found out the negative impact of frequent terms from media on query expansion. This effect reflects the semantic variety of the frequent term connections in texts of different genres.
Fichier principal
Vignette du fichier
ovchinnikova_19103.pdf (193.06 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01913972 , version 1 (06-11-2018)


  • HAL Id : hal-01913972 , version 1
  • OATAO : 19103


Irina Ovchinnikova, Liana Ermakova, Josiane Mothe. Do frequent media words worsen query expansion?. International Scientific Conference Digital Humanities: Resources, Methods, and Research ; a part of the Perm Natural Science forum (2017), May 2017, Perm, Russia. pp. 100-104. ⟨hal-01913972⟩
71 View
26 Download


Gmail Facebook X LinkedIn More