Privately detecting correlations in distributed time series
Sayal, Mehmet; Singh, Lisa
Keyword(s): time-series, correlation, privacy, multi-dimensional
Abstract: In this paper, we consider privacy preservation in the context of independently owned, distributed time series data. Specifically, we are interested in discovering correlations even though we cannot share the raw time series values. We propose developing a generic framework for identifying similarities or correlations of a particular behavior or statistic across participants. Our generic framework makes use of the additive combining property of certain statistics. It also allows for sharing of scaled bin values instead of raw data or statistical values to improve levels of privacy. We find that while there is a natural trade off between privacy and accuracy, we can maintain reasonable correlation accuracy for different levels of privacy.
External Posting Date: October 21, 2010 [Fulltext]. Approved for External Publication
Internal Posting Date: October 21, 2010 [Fulltext]