Microaggregation Algorithm Research Articles

Multivariate Microaggregation of Set-Valued Data

Data controllers manage immense data, and occasionally, it is released publically to help the researchers toconduct their studies. However, this publically shared data may hold personally identifiable information (PII)that can be collected to re-identify a person. Therefore, an effective anonymization mechanism is required toanonymize such data before it is released publically. Microaggregation is one of the Statistical Disclosure Control (SDC) methods that are widely used by many researchers. This method adapts the k-anonymity principle togenerate k-indistinguishable records in the same clusters to preserve the privacy of the individuals. However,in these methods, the size of the clusters is fixed (i.e., k records), and the clusters generated through these methods may hold non-homogeneous records. By considering these issues, we propose an adaptive size clusteringtechnique that aggregates homogeneous records in similar clusters, and the size of the clusters is determinedafter the semantic analysis of the records. To achieve this, we extend the MDAV microaggregation algorithm tosemantically analyze the unstructured records by relying on the taxonomic databases (i.e., WordNet), and thenaggregating them in homogeneous clusters. Furthermore, we propose a distance measure that determines theextent to which the records differ from each other, and based on this, homogeneous adaptive clusters are constructed. In experiments, we measured the cohesiveness of the clusters in order to gauge the homogeneity ofrecords. In addition, a method is proposed to measure information loss caused by the redaction method. In experiments, the results show that the proposed mechanism outperforms the existing state-of-the-art solutions.

Repeated Record Ordering for Constrained Size Clustering

One of the main techniques used in data mining is data clustering, which has many applications in computer science, biology and social sciences. Constrained clustering is a type of clustering in which side information provided by the user is incorporated into current clustering algorithms. One of the well researched constrained clustering algorithms is called microaggregation. In a microaggregation technique, the algorithm divides the dataset into groups containing at least members, where is a user-defined parameter. The main application of microaggregation is in Statistical Disclosure Control (SDC) for privacy preserving data publishing. A microaggregation algorithm is qualified based on the sum of within-group squared error, . Unfortunately, it has been proven that the optimal microaggregation problem is NP-Hard in general, but the special univariate case can be solved optimally in polynomial time. Many heuristics exist for the general case of the problem that are founded on the univariate case. These techniques order multivariate records in a sequence. This paper proposes a novel method for record ordering. Starting from a conventional clustering algorithm, the proposed method repeatedly puts multivariate records into a sequence and then clusters them again. The process is repeated until no improvement is achieved. Extensive experiments have been conducted in this research to confirm the effectiveness of the proposed method for different parameters and datasets.

Mathematically optimized, recursive prepartitioning strategies for k-anonymous microaggregation of large-scale datasets

The technical contents of this work fall within the statistical disclosure control (SDC) field, which concerns the postprocessing of the demographic portion of the statistical results of surveys containing sensitive personal information, in order to effectively safeguard the anonymity of the participating respondents. A widely known technique to solve the problem of protecting the privacy of the respondents involved beyond the mere suppression of their identifiers is the k-anonymous microaggregation. Unfortunately, most microaggregation algorithms that produce competitively low levels of distortions exhibit a superlinear running time, typically scaling with the square of the number of records in the dataset.This work proposes and analyzes an optimized prepartitioning strategy to reduce significantly the running time for the k-anonymous microaggregation algorithm operating on large datasets, with mild loss in data utility with respect to that of MDAV, the underlying method. The optimization strategy is based on prepartitioning a dataset recursively until the desired k-anonymity parameter is achieved. Traditional microaggregation algorithms have quadratic computational complexity in the form Θ(n2). By using the proposed method and fixing the number of recurrent prepartitions we obtain subquadratic complexity in the form Θ(n3/2), Θ(n4/3), ..., depending on the number of prepartitions. Alternatively, fixing the ratio between the size of the microcell and the macrocell on each prepartition, quasilinear complexity in the form Θ(nlog n) is achieved. Our method is readily applicable to large-scale datasets with numerical demographic attributes.

A Micro-Aggregation Algorithm Based on Density Partition Method for Anonymizing Biomedical Data

Objective: Biomedical data can be de-identified via micro-aggregation achieving privacy. However, the existing micro-aggregation algorithms result in low similarity within the equivalence classes, and thus, produce low-utility anonymous data when dealing with a sparse biomedical dataset. To balance data utility and anonymity, we develop a novel microaggregation framework. Methods: Combining a density-based clustering method and classical micro-aggregation algorithm, we propose a density-based second division micro-aggregation framework called DBTP . The framework allows the anonymous sets to achieve the optimal k- partition with an increased homogeneity of the tuples in the equivalence class. Based on the proposed framework, we propose a k − anonymity algorithm DBTP − MDAV and an l − diversity algorithm DBTP − l − MDAV to respond to different attacks. Conclusion: Experiments on real-life biomedical datasets confirm that the anonymous algorithms under the framework developed in this paper are superior to the existing algorithms for achieving high utility.

Improving k-anonymity based privacy preservation for collaborative filtering

Collaborative Filtering (CF) is applied in recommender systems to predict users’ preference through filtering the information or patterns. Privacy Preserving Collaborative Filtering (PPCF) aims to achieve privacy protection in the recommendation process, which has an increasing significance in recommender systems and thus attracted much interests in recent years. Existing PPCF methods are mainly based on cryptography, obfuscation, perturbation and differential privacy. They have high computational cost, low data quality and difficulties in calibrating the magnitude of noise. This paper proposes a (p, l, α)-diversity method that improves the existing k-anonymity method in PPCF, where p is attacker’s prior knowledge about users’ ratings and (l, α) is the diversity among users in each group to improve the level of privacy preserving. To achieve (l, α)-diversity, users in each equivalence class shall come from at least l (l < k) clusters in α clustering results. Therefore, we firstly apply Latent Factor Model (LFM) to reduce matrix sparsity. Then we propose an improved Maximum Distance to Average Vector (MDAV) microaggregation algorithm based on importance partitioning to increase the homogeneity among the records in each group which can retain better data quality in (p, l, α)-diversity model. Finally, we apply t-closeness in PPCF. Theoretical analysis and experimental results demonstrate our approach assures a higher level of privacy preserving and less information loss than existing methods.

E-DMDAV: A new privacy preserving algorithm for wearable enterprise information systems

ABSTRACTWearable devices have been widely used in many fields to improve the quality of people’s lives. More and more data on individuals and businesses are collected by statistical organizations though those devices. Almost all of this data holds confidential information. Statistical Disclosure Control (SDC) seeks to protect statistical data in such a way that it can be released without giving away confidential information that can be linked to specific individuals or entities. The MDAV (Maximum Distance to Average Vector) algorithm is an efficient micro-aggregation algorithm belonging to SDC. However, the MDAV algorithm cannot survive homogeneity and background knowledge attacks because it was designed for static numerical data. This paper proposes a systematic dynamic-updating anonymity algorithm based on MDAV called the e-DMDAV algorithm. This algorithm introduces a new parameter and a table to ensure that the k records in one cluster with the range of the distinct values in each cluster is no less than e for numerical and non-numerical datasets. This new algorithm has been evaluated and compared with the MDAV algorithm. The simulation results show that the new algorithm outperforms MDAV in terms of minimizing distortion and disclosure risk with a similar computational cost.

Enhancing aggregation phase of microaggregation methods for interval disclosure risk minimization

Microaggregation is a masking mechanism to protect confidential data in a public release. This technique can produce a k-anonymous dataset where data records are partitioned into groups of at least k members. In each group, a representative centroid is computed by aggregating the group members and is published instead of the original records. In a conventional microaggregation algorithm, the centroids are computed based on simple arithmetic mean of group members. This naive formulation does not consider the proximity of the published values to the original ones, so an intruder may be able to guess the original values. This paper proposes a disclosure-aware aggregation model, where published values are computed in a given distance from the original ones to attain a more protected and useful published dataset. Empirical results show the superiority of the proposed method in achieving a better trade-off point between disclosure risk and information loss in comparison with other similar anonymization techniques.

Fast data-oriented microaggregation algorithm for large numerical datasets

Microaggregation is a successful mechanism to solve the tension between respondent privacy and data quality in the context of Statistical Disclosure Control. Microaggregation, for numerical datasets, is defined as a clustering problem with the constraint of having at least k records in each group, such that the sum of the within-group squared error (SSE) is minimized. Unfortunately, the data publisher has to execute an algorithm iteratively for different values of k to investigate a good trade-off between privacy and utility. Multiple execution of an algorithm on large numerical datasets is resource wasting, since most of the computations are repetitive. In this paper, we propose a Fast Data-oriented Microaggregation algorithm (FDM) that efficiently anonymizes large multivariate numerical datasets for multiple successive values of k. Experimental results on real world datasets demonstrate the superiority of the method in terms of both the data quality and time complexity. Moreover, the method usually achieves a better trade-off between disclosure risk and information loss of the protected dataset in comparison with previous techniques.

A modification of the Lloyd algorithm for k-anonymous quantization

We address the problem of designing quantizers that cluster data while satisfying a k-anonymity requirement. A general data compression perspective is adopted, which considers both discrete and continuous probability distributions, and corresponding constraints on both cell sizes and quantizer index probabilities. Potential applications of this problem extend well beyond the important case of microdata anonymization, to include also optimized task allocation under workload constraints. Our contribution is twofold. First and most importantly, we present a theoretical analysis showing the optimality conditions which probability-constrained quantizers must satisfy, thereby theoretically characterizing optimal k-anonymous aggregation as a special case. As a second contribution, inspired by our theoretical analysis, we propose an alternating optimization algorithm for the design of this type of quantizers. Our algorithm is conceptually motivated by the popular Lloyd–Max algorithm for quantization design, originally intended for data compression, also known as the k-means method. Experimental results for synthetic and real data, with mean squared error as a distortion measure, confirm that our method outperforms MDAV, a popular fixed-size microaggregation algorithm for statistical disclosure control. This performance improvement is in terms of data utility, for the exact same k-anonymity constraint, but does come at the expense of higher computational sophistication.

An algorithm for k-anonymous microaggregation and clustering inspired by the design of distortion-optimized quantizers

We present a multidisciplinary solution to the problems of anonymous microaggregation and clustering, illustrated with two applications, namely privacy protection in databases, and private retrieval of location-based information. Our solution is perturbative, is based on the same privacy criterion used in microdata k-anonymization, and provides anonymity through a substantial modification of the Lloyd algorithm, a celebrated quantization design algorithm, endowed with numerical optimization techniques.Our algorithm is particularly suited to the important problem of k-anonymous microaggregation of databases, with a small integer k representing the number of individual respondents indistinguishable from each other in the published database. Our algorithm also exhibits excellent performance in the problem of clustering or macroaggregation, where k may take on arbitrarily large values. We illustrate its applicability in this second, somewhat less common case, by means of an example of location-based services. Specifically, location-aware devices entrust a third party with accurate location information. This party then uses our algorithm to create distortion-optimized, size-constrained clusters, where k nearby devices share a common centroid location, which may be regarded as a distorted version of the original one. The centroid location is sent back to the devices, which use it when contacting untrusted location-based information providers, in lieu of the exact home location, to enforce k-anonymity.We compare the performance of our novel algorithm to the state-of-the-art microaggregation algorithm MDAV, on both synthetic and standardized real data, which encompass the cases of small and large values of k. The most promising aspect of our proposed algorithm is its capability to maintain the same k-anonymity constraint, while outperforming MDAV by a significant reduction in data distortion, in all the cases considered.

TFRP: An efficient microaggregation algorithm for statistical disclosure control

Recently, the issue of statistic disclosure control (SDC) has attracted much attention. SDC is a very important part of data security dealing with the protection of databases. Microaggregation for SDC techniques is widely used to protect confidentiality in statistical databases released for public use. The basic problem of microaggregation is that similar records are clustered into groups, and each group contains at least k records to prevent disclosure of individual information, where k is a pre-defined security threshold. For a certain k, an optimal multivariable microaggregation has the lowest information loss. The minimum information loss is an NP-hard problem. Existing fixed-size techniques can obtain a low information loss with O( n 2) or O( n 3/ k) time complexity. To improve the execution time and lower information loss, this study proposes the Two Fixed Reference Points (TFRP) method, a two-phase algorithm for microaggregation. In the first phase, TFRP employs the pre-computing and median-of-medians techniques to efficiently shorten its running time to O( n 2/ k). To decrease information loss in the second phase, TFRP generates variable-size groups by removing the lower homogenous groups. Experimental results reveal that the proposed method is significantly faster than the Diameter and the Centroid methods. Running on several test datasets, TFRP also significantly reduces information loss, particularly in sparse datasets with a large k.