S i ² p

Abstract

The incomplete data is universal in many real-life applications due to data integration, the limitation of devices, etc. In this demonstration, we present S i 2 p , a restaurant recommendation <u>S</u>ystem with <u>P</u>reference queries on <u>I</u>ncomplete <u>I</u>nformation. S i 2 p is capable of friendly recommending desirable restaurants based on preference queries that take the incomplete ratings information into consideration. It adopts the browser-server model, and incorporates three functionality modules including friendly and convenient query submission , flexible and useful result explanation , timely and incremental dataset interaction. S i 2 p provides the server side based on an extended PostgreSQL database that integrates two types of preference queries, namely, skyline and top- k dominating queries over incomplete data. It also offers the browser-based interface for the users to interact with the system. Using a real restaurant dataset from TripAdvisor, we demonstrate S i 2 p can recommend and explore the restaurants in a friendly way.