Empirical Evaluation of Explicit versus Implicit Acquisition of User Profiles in Information Filtering Systems

摘要

Three different approaches of profile acquisition in a filtering system were compared. The explicit mode allowed users to directly specify the profile, the implicit mode only permitted relevance feedback, and the combined mode allowed users to initialize the profile using explicit means and refine it continuously using relevance feedback. The SIFTER system, designed by the Indiana University Filtering Research Group, was used to conduct a series of experiments. A collection 6000 records retrieved from the Ebsco Health Sources Plus database was the test data set. Eighteen subjects were randomly assigned to three different versions of the system (three experimental levels), and each subject performed 15 sessions. Filtering performance measured using normalized precision showed that the three approaches were significantly different (α=0.05 & p=0.012). Overall, it was found that the explicit means of profile acquisition consistently produced superior results. The findings also showed that exclusive reliance on relevance feedback in the implicit mode resulted in inferior performance.