TREC 2012 Microblog Track Experiments at Kobe University

Abstract

This paper describes our approach to real-time ad hoc task processing in the TREC 2012 Microblog track. The approach uses two-stage relevance feedback to model search interests and temporal dynamics on a microblog. The first relevance feedback uses a single user-selected tweet as feedback. The second approach uses time-based query expansion method leveraging the temporal property derived from the real-time feature on microblogging services. The experimentally obtained results demonstrate that our two-stage relevance feedback approaches improve search result relevance considerably.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 2012
Accession Number
ADA581249

Entities

People

  • Kazuhiro Seki
  • Kuniaki Uehara
  • Taiki Miyanishi

Organizations

  • Kobe University

Tags

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Detectors
  • Equations
  • Feedback
  • Information Operations
  • Information Retrieval
  • Language
  • Learning
  • Precision
  • Schools
  • Standards
  • Test And Evaluation
  • Universities

Fields of Study

  • Computer science

Readers

  • Computational Modeling and Simulation
  • Information Retrieval