SIRA: TREC Session Track 2014

Abstract

This paper discusses Siena s Interactive Research Assistant s (SIRA) participation in the Text Retrieval Conference (TREC) Session Track of 2014. The overall goal of this track is to improve search results during query sessions based on a user s behavior. Query sessions include many aspects of a search, including query topics, initial retrieved webpages, clicked on links, visit times, etc. SIRA has used several methods to improve search results that will be discussed in this paper. Each method of query expansion utilized clicked-on and non-clicked-on links, pages with the longest visited time, and N-Percent (N%) of each page. Two of our three submissions improved over our baseline results and both of these were equal to the median submission for all participants in the track.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 2014
Accession Number
ADA618621

Entities

People

  • Brennan Bushee
  • Drew Pintus
  • Patrick T Smith
  • Sharon G. Small

Organizations

  • Siena College

Tags

Communities of Interest

  • Ground and Sea Platforms
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Abstracts
  • Artificial Intelligence
  • Availability
  • Department Of Defense
  • Frequency
  • Information Operations
  • Information Retrieval
  • Intelligence Analysts
  • Navy
  • Nuclear Powered Submarines
  • Standards
  • Submarines
  • Test And Evaluation
  • Universities
  • Weapons Of Mass Destruction
  • Word Lists

Fields of Study

  • Computer science

Readers

  • Information Retrieval