Adaptive Suggestions for Example based Tools

Preference-based search is the problem of finding an item that matches best with a user’s preferences. Recently several researchers have proposed approaches based on examples where preferences are stated in form of critiques on shown options. User studies show that example-based tools for preference-based search can achieve significantly higher accuracy when they are complemented with suggestions to stimulate preference expression. This paper proposes a method for producing adaptive suggestions taking in consideration prior distribution of preferences (perhaps learned from a population of similar people) and reactions of the user to the shown examples. At each interaction cycle, a set of best matches (candidate options) and suggestions are retrieved and shown to the user. Using Bayes rule, the distribution is updated at each step. We evaluate the approach with simulations.