Experimental evaluation of query fingerprinting with embedded search term markers

  • Authors:

  • Beomjin Kim

  • Affiliations:

  • Department of Computer Science, Indiana University-Purdue University, Fort Wayne, IN

  • Venue:
  • ICCSA'03 Proceedings of the 2003 international conference on Computational science and its applications: PartII
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Previous studies have shown that visualization is an effective solution in finding relevant information from a huge collection of information. This paper presents experimental results of an information visualization system, which presents retrieved articles through visual abstraction. In order to address the difficulties in interpreting the color codes, the Query Fingerprinting system has been reinforced with embedded search term markings on the color codes. For a non-biased evaluation of the developed algorithms, a series of experiments has been executed on TREC-5 text collection with their corresponding queries and relevance judgments. The results show that the participants who utilized the developed method in judging the relevancy of articles showed better performance while maintaining the accuracy than the users who utilized a system based on the traditional approach. The participants also have shown a high degree of satisfaction with the search term markings in understanding the distributional patterns of queries. The experiment acknowledges the feasibility of the Query Fingerprinting with embedded search term markers as a search-supporting tool, which assists users' decision-making processes and lessens the search activities.