In this system, a special type of the user model is supported called as a task model. It is aimed at supporting information analysts who are required to solve a specific task. A very short term user model is constructed to model each different user task and personalization is made with the task models. A post filtering approach was applied to this purpose. The ad-hoc retrieval results triggered by user queries were re-ordered using the task model so that more relevant documents to the task model should be ranked up in the lists. In terms of the interface, we added a control with which users can control the effects of the query and the task model according to their needs and the context.
System development was currently supported by DARPA GALE project
Follow this link to try TaskSieve
J.-W. Ahn, P. Brusilovsky, D. He, J. Grady, and Q. Li, "Personalized web exploration with task models," in WWW '08: Proceeding of the 17th international conference on World Wide Web. New York, NY, USA: ACM, 2008, pp. 1-10.
Jae-wook Ahn
|