Clustering of multi-source data

Due to the advent of pervasive and distributed computing, the shift in data processing paradigms is quite visible in many applications (e.g., communication, industrial control, public administrations, geographically distributed business, etc.). The switch from centralized processing paradigm to the local processing paradigm is becoming a necessity due to many factors such as: security, privacy, limitation of the transmission bandwidth, time processing, business strategies and rules, limited power supply, etc. In this paper, a clustering approach dedicated to distributed data emanating from diverse sources is proposed and evaluated. This approach is guided by two principles: coordination and collaboration among various data sources to cope with the problem of privacy and corporate confidentiality in an appropriate way.