Additive Sanitization: A Technique for Pattern-Preserving Anonymization for Time-Series Data

A time series is a set of data normally collected at usual intervals and often contains huge amount of individual privacy. The need to protect privacy and anonymization of time-series while trying to support complex queries such as pattern range and pattern matching queries. The conventional (k, p)-anonymity model cannot effectively address this problem as it may suffer serious pattern loss. In the proposed work a new technique called additive sanitization has been developed which increment the supports of item sets and their subsets in order to reduce pattern loss and prevent linkage attack.