Putting Statistical Disclosure Control into Practice: The ARX Data Anonymization Tool

Praßer, Fabian; Kohlmayer, Florian

doi:10.1007/978-3-319-23633-9_6

Cited by 64 publications

(61 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The software described in this article has been developed by extending ARX, an open source anonymization tool which has specifically been designed for applications to biomedical data [19]. In this section, we will focus on the two most important functionalities implemented, which are (1) methods for the automated creation of privacypreserving prediction models and (2) methods for evaluating and fine-tuning the resulting models.…”

Section: Methodsmentioning

confidence: 99%

“…In practice, ARX implements a wide range of pruning strategies and optimization techniques to avoid needing to analyze all possible output datasets (see, e.g. [19,21]). Moreover, ARX supports further transformation techniques which are implemented by extending the basic anonymization process outlined in this paragraph.…”

Section: Methods For Creating Privacy-preserving Prediction Modelsmentioning

confidence: 99%

“…In prior work, we have already described a basic implementation of the approach [18]. However, the implementation had several important limitations, which resulted from the compressed internal data representation used by ARX [19]: (1) it only supported one class variable, (2) it required that class variables were addressed by a privacy model, and (3) it required that no transformations were applied to target variables. To Fig.…”

Section: Utility Modelsmentioning

confidence: 99%

See 2 more Smart Citations

A comprehensive tool for creating and evaluating privacy-preserving biomedical prediction models

Eicher

Bild

Spengler

et al. 2020

BMC Med Inform Decis Mak

Self Cite

View full text Add to dashboard Cite

Background: Modern data driven medical research promises to provide new insights into the development and course of disease and to enable novel methods of clinical decision support. To realize this, machine learning models can be trained to make predictions from clinical, paraclinical and biomolecular data. In this process, privacy protection and regulatory requirements need careful consideration, as the resulting models may leak sensitive personal information. To counter this threat, a wide range of methods for integrating machine learning with formal methods of privacy protection have been proposed. However, there is a significant lack of practical tools to create and evaluate such privacy-preserving models. In this software article, we report on our ongoing efforts to bridge this gap. Results: We have extended the well-known ARX anonymization tool for biomedical data with machine learning techniques to support the creation of privacy-preserving prediction models. Our methods are particularly well suited for applications in biomedicine, as they preserve the truthfulness of data (e.g. no noise is added) and they are intuitive and relatively easy to explain to non-experts. Moreover, our implementation is highly versatile, as it supports binomial and multinomial target variables, different types of prediction models and a wide range of privacy protection techniques. All methods have been integrated into a sound framework that supports the creation, evaluation and refinement of models through intuitive graphical user interfaces. To demonstrate the broad applicability of our solution, we present three case studies in which we created and evaluated different types of privacy-preserving prediction models for breast cancer diagnosis, diagnosis of acute inflammation of the urinary system and prediction of the contraceptive method used by women. In this process, we also used a wide range of different privacy models (k-anonymity, differential privacy and a game-theoretic approach) as well as different data transformation techniques. Conclusions: With the tool presented in this article, accurate prediction models can be created that preserve the privacy of individuals represented in the training set in a variety of threat scenarios. Our implementation is available as open source software.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methods For Creating Privacy-preserving Prediction Modelsmentioning

confidence: 99%

Section: Utility Modelsmentioning

confidence: 99%

See 1 more Smart Citation

A comprehensive tool for creating and evaluating privacy-preserving biomedical prediction models

Eicher

Bild

Spengler

et al. 2020

BMC Med Inform Decis Mak

Self Cite

View full text Add to dashboard Cite

show abstract

“…These measures are generally divided into two categories: (1) entropy-based and (2) distance-based (e.g., the Hellinger distance). Furthermore, most of the above mentioned usefulness quantification approaches are implemented into data anonymization tools, such as ARX [22].…”

Section: Introductionmentioning

confidence: 99%

Assessing Data Usefulness for Failure Analysis in Anonymized System Logs

Ghiasvand¹,

Ciorba²

2018

2018 17th International Symposium on Parallel and Distributed Computing (ISPDC)

View full text Add to dashboard Cite

System logs are a valuable source of information for the analysis and understanding of systems behavior for the purpose of improving their performance. Such logs contain various types of information, including sensitive information. Information deemed sensitive can either directly be extracted from system log entries by correlation of several log entries, or can be inferred from the combination of the (non-sensitive) information contained within system logs with other logs and/or additional datasets. The analysis of system logs containing sensitive information compromises data privacy. Therefore, various anonymization techniques, such as generalization and suppression have been employed, over the years, by data and computing centers to protect the privacy of their users, their data, and the system as a whole. Privacy-preserving data resulting from anonymization via generalization and suppression may lead to significantly decreased data usefulness, thus, hindering the intended analysis for understanding the system behavior. Maintaining a balance between data usefulness and privacy preservation, therefore, remains an open and important challenge. Irreversible encoding of system logs using collision-resistant hashing algorithms, such as SHAKE-128, is a novel approach previously introduced by the authors to mitigate data privacy concerns. The present work describes a study of the applicability of the encoding approach from earlier work on the system logs of a production high performance computing system. Moreover, a metric is introduced to assess the data usefulness of the anonymized system logs to detect and identify the failures encountered in the system. 1 System log and syslog are used interchangeably in this work 2 https://top500.org/list/2017/11/ arXiv:1805.01790v1 [cs.DC] 4 May 2018Germany is conducted. Moreover, a metric is introduced to assess the usefulness of data within system logs during and postencoding for the purpose of detecting and identifying failures encountered in the HPC system. The use of the proposed metric on the considered system logs show that the anonymized data retains a high degree of usefulness for the intended purpose of failure detection and identification.The novelty and contributions of the anonymization approach to retain data usefulness proposed in this work consists of:(1) anonymization of unstructured system log messages; (2) preservation of the usefulness of system logs (especially for semi-/automated failure analysis); (3) a guarantee for protecting data privacy; (4) significant reduction of the required capacity for the storage of the anonymized system logs; (5) offering the choice to set the degree of generalization according to the intended data usage; and (6) providing readily analyzable data, anonymized and encoded, that does not require decoding before analysis.The remainder of this work is structured as follows. Section II overviews the work related to data anonymization and usefulness. The anonymization approach employed herein is described in Section III. The proposed...

show abstract

“…There are a number of tools available to anonymise data, which also provide some risk analysis feedback. The ARX Tool [10] provides methods for analyzing re-identification risks following the prosecutor, journalist and marketer attacker models on a number of anonymisation algorithms. The Cornell Anonymization Toolkit (CAT) [11] performs Risk Analysis in terms evaluating the disclosure of risks of each record in anonymised data based on user specified assumptions about the adversarys background knowledge.…”

Section: Related Workmentioning

confidence: 99%

Identifying Privacy Risks in Distributed Data Services: A Model-Driven Approach

Grace

Burns

Neumann

et al. 2018

2018 IEEE 38th International Conference on Distributed Computing Systems (ICDCS)

View full text Add to dashboard Cite

Abstract-Online services are becoming increasingly datacentric; they collect, process, analyze and anonymously disclose growing amounts of personal data. It is crucial that such systems are engineered in a privacy-aware manner in order to satisfy both the the privacy requirements of the user, and the legal privacy regulations that the system operates under. How can system developers be better supported to create privacy-aware systems and help them to understand and identify privacy risks? Model-Driven Engineering (MDE) offers a principled approach to engineer systems software. The capture of shared domain knowledge in models and corresponding tool support can increase the developers' understanding. In this paper, we argue for the application of MDE approaches to engineer privacyaware systems. We present a general purpose privacy model and methodology that can be used to analyse and identify privacy risks in systems that comprise both access control and data pseudonymization enforcement technologies. We evaluate this method using a case-study based approach and show how the model can be applied to engineer privacy-aware systems and privacy policies that reduce the risk of unintended disclosure.

show abstract

Putting Statistical Disclosure Control into Practice: The ARX Data Anonymization Tool

Cited by 64 publications

References 29 publications

A comprehensive tool for creating and evaluating privacy-preserving biomedical prediction models

A comprehensive tool for creating and evaluating privacy-preserving biomedical prediction models

Assessing Data Usefulness for Failure Analysis in Anonymized System Logs

Identifying Privacy Risks in Distributed Data Services: A Model-Driven Approach

Contact Info

Product

Resources

About