Consumer Data and Privacy in Ubiquitous Computing: Master's thesis

The privacy-preserving perspective on data mining is relatively young area. The research done in the area is mainly theoretical. The current trend with growing amount of personalization in online services has created also applications for personalized marketing. Personalized marketing services use detailed information about the context and personal history of a customer. This needs sophisticated individual identification methods, which raise privacy concern. The novelty in privacy-preserving methods is that sensitive and distributed data could be used for data mining task and the privacy of individuals is preserved. This work has two objectives: first is to use consumer data from distributed sources and study how customer segmentation is possible while preserving the privacy. The idea is to conduct the customer segmentation in a way that the data need not leave the agent holding the data. The other objective is the value of the knowledge acquired from collectively conducted segmentation. We believe that collectively conducted segmentation produces knowledge that cannot be acquired otherwise. The results of this work show that privacypreserving customer segmentation is possible and the collectively conducted segmentation produces new knowledge.