Reverse Top-k Query on Uncertain Data

Traditional top-k query returns the products to customers according to their preferences,whereas reverse top-k query returns the preferences for which a given product is in the top-k result.Reverse top-k query is valuable in business analysis because it can help the manufacturers evaluate the impact of a product on customers.Existing reverse top-k query assumes the underlying data is certain,however,uncertainty arises in many real applications.In this paper,we extended reverse top-k query to uncertain data,and provided an efficient query algorithm named GMV based on materialized views.Our experimental evaluation demonstrates that,GMV can reduce the preferences which need to be computed,and achieves relatively high computational efficiency.