Time-based analysis of search data logs


Ozmutlu H. C., Spink A.

2nd International Conference on Internet Computing (IC 2001), Nevada, United States Of America, 25 - 28 June 2001, pp.41-46 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • City: Nevada
  • Country: United States Of America
  • Page Numbers: pp.41-46
  • Bursa Uludag University Affiliated: Yes

Abstract

A common way to retrieve information from the Web is through search engines. Understanding Web users' information seeking behavior is very important in developing efficient search engines. The main purpose of our study is to perform a time-based analysis of Web users' search behavior to aid the development of intelligent user based search engines. Web query sessions used in this study are filtered from a query log provided by the Excite search engine. We found that the number and character of queries submitted to a Web search engine changes with respect to the hour of the day. Results show that the mean number of queries per session, number of queries and user sessions, and the duration of queries and user sessions declined in a near linear fashion (anywhere from 42% to 91%) as the workday progresses. Understanding such time-based query load and user information retrieval needs is useful for search engines reallocating of resources and efficient operations.