Reverse Top-k Query Result Analysis and Refinement System

Abstract

Reverse top-k query is an important tool for market analysis since it can help manufacturers to identify potential customers for their products. Unfortunately, users may get unexpected query results after performing the reverse top-k query. There are two types of unexpected query results that are of interest in this demonstration: (i) content-based unexpected query results, (ii) cardinality-based unexpected query results. Toward this, we develop IS2R, an interactive system that can refine the reverse top-k query to eliminate the unexpected query results. After refining, IS2R guarantees that the expected (unexpected) objects appear (disappear); and/or the cardinality of query result will satisfy user’s requirements. The IS2R returns the refinement suggestions with the minimal cost which hinge on the predefined penalty models. In this demonstration, we show different scenarios on how IS2R can be used to analyze and refine the original reverse top-k query and explore its effectiveness and efficiency.