The aim of a Personalised Ranking Model (PRM) is to filter the topk set of documents from a number of relevant documents matching the search query. Dwell times of previously clicked results have been shown to be valuable for estimating documents’ relevance. The indexing structure of the dwell time is an important parameter. We propose a dwell time-based scoring scheme called Dwell-tf-idf to index text and non-text data, based on which search results are ranked. The effectiveness of incorporating into the ranking process the proposed Dwell-tf-idf scheme is validated by a controlled experiment which shows a significant improvement in the search results within the top-k rank.
CITATION STYLE
Al-Sharji, S., Beer, M., & Uruchurtu, E. (2015). A dwell time-based technique for personalised ranking model. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9262, pp. 205–214). Springer Verlag. https://doi.org/10.1007/978-3-319-22852-5_18
Mendeley helps you to discover research relevant for your work.