본문 바로가기
  • Home

Methodology for Search Intent-based Document Recommendation

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2021, 26(6), pp.115-127
  • DOI : 10.9708/jksci.2021.26.06.115
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : May 6, 2021
  • Accepted : June 24, 2021
  • Published : June 30, 2021

LEE DONG HOON 1 Namgyu Kim 1

1국민대학교

Accredited

ABSTRACT

It is not an easy task for a user to find the correct documents that a user really wanted at once from a vast amount of the search results. For this reason, various methods of recommending documents by taking the user’s preferences into consideration based on the user's document browsing history have been proposed. However, the document recommendation methodology based on the document browsing history also has a limitation that only the information the user has viewed is utilized, but the intent of the user searching for the document is not fully utilized. Therefore, we propose a document recommendation method based on the user’s search intent that utilizes information on “Why” the user reads the document, instead of the information on “Who” reads the document. In order to confirm the feasibility of the proposed methodology, an experiment was conducted by analyzing 239,438 actual user’s search history of one of the most popular e-commerce platform companies in Korea. As a result, our methodology showed superior performance compared to the existing content-based or simple browsing history-based recommendation model.

Citation status

* References for papers published after 2022 are currently being built.