Privacy-Aware Data Publishing and Integration for Collaborative Service Recommendation

Abstract

Item-based collaborative filtering (i.e., ICF) technique has been widely recruited to make service recommendations in the big data environment. However, the ICF technique only performs well when the data for service recommendation decision-making are stored in a physically centralized manner, while they often fail to recommend appropriate services to a target user in the distributed environment where the involved multiple parties are reluctant to release their data to each other due to privacy concerns. Considering this drawback, we improve the traditional ICF approach by integrating the locality-sensitive hashing (LSH) technique, to realize secure and reliable data publishing. Furthermore, through integrating the published data with little privacy across different platforms, appropriate services are recommended based on our suggested recommendation approach named ICF <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LSH</sub> . At last, simulated experiments are conducted on WS-DREAM data set. Experiment results show that ICF <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LSH</sub> performs better than the competitive approaches in terms of service recommendation accuracy, efficiency, and the capability of privacy-preservation.