ABSTRACT
In Chapter 2, we covered privacy preservation of multidimensional data, which is the most common form of data structure found in any enterprise. As seen in Chapter 2, a rich set of privacy preserving algorithms are currently in use. Emphasis on deriving knowledge from data along with the rapid evolution of digital platforms in enterprises have resulted in more and more complex data structures being analyzed. Nowadays, it is common to find complex data structures such as graph data, time series data, longitudinal data, transaction data, and spatiotemporal data in many enterprises. Mining these complex data results in a lot of useful insights that can benefit businesses. But these complex data structures contain personally identifiable information that must be anonymized before they can be mined. Multidimensional data, which were studied in the previous chapters, are much simpler in their structure, and traditional anonymization techniques can be used on such data. But these techniques cannot be used for complex data structures. The more complex the data structure, the more the avenues are to be re-identified. Therefore, anonymization techniques that address the various dimensions of a data structure are required. In this part, we examine the anonymization techniques for graph data, time series data, longitudinal data, and transaction data.
