COCOA: A Synthetic Data Generator for Testing Anonymization Techniques

Abstract

Abstract. Conducting extensive testing of anonymization techniques is critical to assess their robustness and identify the scenarios where they are most suitable. However, the access to real microdata is highly restricted and the one that is publicly-available is usually anonymized or aggregated; hence, reducing its value for testing purposes. In this paper, we present a framework (COCOA) for the generation of realistic synthetic microdata that allows to define multi-attribute relationships in order to preserve the functional dependencies of the data. We prove how COCOA is useful to strengthen the testing of anonymization techniques by broadening the number and diversity of the test scenarios. Results also show how COCOA is practical to generate large datasets.

A unified approach to automate the usage of plagiarism detection tools in programming courses

“…For instance, the work on [6] presents a semiautomated approach to facilitate the grading of programming code structures, while the work on [7] presents an automated approach to protect sensitive information before sharing it with third parties. Similarly, other works have focused on automatically generating testing data [8], evaluating pedagogical e-learning content [9], or constructing generalisation hierarchies to anonymise categorical data [10]. In the particular case of plagiarism detection of source code, several tools have been developed in order to support users in this task [11].…”

In-Test Adaptation of Workload in Enterprise Application Performance Testing

“…Moreover, the work on [13] introduced an approach to assess the performance of a distributed memory program in a clustered environment. Meanwhile, other works have focused on generating realistic testing data [9], or providing techniques to facilitate the monitoring of performance counters [21]. Finally, other efforts have also centred on reducing the expertise required.…”

One Size Does Not Fit All

“…Similarly, the work on [23] presents a technique to identify the early warning signs that typically precede a relevant performance degradation in a system. Moreover, some other works have centred on generating useful synthetic testing data [10,11], or on providing techniques that can reduce the expertise required in order to efficiently automate the usage of the diagnosis tools in the performance testing domain [20]. In contrast to these works, which aim to improve other facets of performance testing, our solution addresses the particular need of setting a suitable test workload for a particular application; hence, successfully isolating a user from the complexities of determining such workload.…”