Re-ranking the Search Results for Users with Time-periodic Intents

Horscroft, Gina and Nagar, Jivashi and Suleman, Hussein (2018) Re-ranking the Search Results for Users with Time-periodic Intents, Proceedings of 2018 Annual Conference of the South African Institute of Computer Scientists and Information Technologists (SAICSIT), 26-28 September 2018, Port Elizabeth, ACM.

[img] PDF
Re_ranking_for_users_with_distinct_intents.pdf

Download (836kB)

Abstract

This paper investigates the time of search as a feature to improve the personalization of information retrieval systems. In general, users issue small and ambiguous queries, which can refer to different topics of interest. Although personalized information retrieval systems take care of user’s topics of interest, but they do not consider if the topics are time periodic. The same ranked list cannot satisfy user search intents every time. This paper proposes a solution to rerank the search results for time sensitive ambiguous queries. An algorithm "HighTime" is presented here to disambiguate the time sensitive ambiguous queries and re-rank the default Google results by using a time sensitive user profile. The algorithm is evaluated by using two comparative measures, MAP and NDCG. Results from user experiments showed that re-ranking of search results based on HighTime is effective in presenting relevant results to the users.

Item Type: Conference paper
Subjects: Information systems > Information retrieval
Date Deposited: 09 Nov 2018
Last Modified: 10 Oct 2019 15:31
URI: http://pubs.cs.uct.ac.za/id/eprint/1290

Actions (login required)

View Item View Item