Mining myself in the community: Privacy preserved crowd sensing and computing

Abstract

With the raising popularity of online/mobile social applications, many individuals are increasingly attracted to their relative positions when compared to others in terms of emotional mood, travelling location, walking distance, fitness status, etc. These interest can be summarized as one question “where am I in my community?”. However, it often forms a deadlock that people are interested in the others’ data but are unwilling to disclose their own information (mood, health, etc.). In order to break the deadlock, we propose a privacy preserving participatory sensing scheme that will not disclose individual’s privacy. Specifically, we present a privacy preservation data gathering approach and adopt an improved data mining algorithm to acquire a polynomial approximation function model on distributed user data to provide a privacy preservation method in participatory sensing. Experiments demonstrate that our approach can achieve a valid result comparing with the result without privacy preservation.