Personalizing Information Retrieval Using Task Stage and Task Type
Jun Liu · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · 2010
This paper reports the findings on the roles that task stage and task type play in predicting the usefulness of documents from time that users spent on them. A 3-stage controlled lab experiment was conducted with 24 participants, each working on 3 sub-tasks in a general task, couched either as parallel or dependent. They were asked to search for information and write a report on the topic. Results showed that task stage can enhance the accuracy of predicting document usefulness from decision time, especially in the parallel task. These findings are extremely helpful in predicting document usefulness and accordingly in tailoring search towards specific users and user situation.