TítolA Cost-Continuity Model for Web Search
Publication TypeBook Chapter
Year of Publication2010
AuthorsNettleton D, Codina-Filba J
Book TitleLecture Notes in Artificial Intelligence
Paraules claudata analysis and modeling, web query-sessions, Web-mining

In this paper we present and empirically evaluate a ‘continuity-cost model’ for Internet query sessions made by users. We study the relation of different ‘cost factors’ for a user query session, with the continuity of the user in that query session, and the order of the query in the query session. We define cost indicators from the available query log data, which are to be studied in relation to continuity and to the order/number of the query (1st, 2nd, 3rd, ..). One of our hypotheses is that cost related factors will reflect the step by step nature of the query session process. We use descriptive statistics together with rule induction to identify the most relevant factors and observable trends, and produce three classifier data models, one for each ‘query number’, using the ‘continuity flag’ as classifier label. Using the cost factors, we identify trends relating continuity/query number to user behavior, and we can use that information, for example, to make decisions about caching and query recommendation.