A Multi-Level Privacy-Preserving Approach to Hierarchical Data Based on Fuzzy Set Theory

Abstract

Nowadays, more and more applications are dependent on storage and management of semi-structured information. For scientific research and knowledge-based decision-making, such data often needs to be published, e.g., medical data is released to implement a computer-assisted clinical decision support system. Since this data contains individuals’ privacy, they must be appropriately anonymized before to be released. However, the existing anonymization method based on l-diversity for hierarchical data may cause a serious similarity attack, and cannot protect data privacy very well. In this paper, we utilize fuzzy sets to divide levels for sensitive numerical and categorical attribute values uniformly (a categorical attribute value can be converted into a numerical attribute value according to its frequency of occurrences), and then transform the value levels to sensitivity levels. The privacy model ( α l e v h , k)-anonymity for hierarchical data with multi-level sensitivity is proposed. Furthermore, we design a privacy-preserving approach to achieve this privacy model. Experiment results demonstrate that our approach is obviously superior to existing anonymous approach in hierarchical data in terms of utility and security.

Proximity Measurement for Hierarchical Categorical Attributes in Big Data

“…Otherwise, Wang et al [10] suggested a multilevel privacy-preserving approach based on fuzzy sets to ensure privacy. e proposed algorithm treats both numerical and categorical sensitive attributes by converting the categorical attribute value into a numerical attribute value based on its occurrence frequency.…”

“…e proposed algorithm treats both numerical and categorical sensitive attributes by converting the categorical attribute value into a numerical attribute value based on its occurrence frequency. Unlike the previous algorithms, the suggested one in [10] can resist the similarity attack by partitioning sensitive values into five levels and then setting a sensitivity level for each sensitive value. Based on semantic rules, Mubark et al [39] proposed a technique dealing with categorical data.…”

“…Our proposed technique resists the similarity attack by using a simpler method compared to the suggested techniques offering resistance to the similarity attack already mentioned. We do not need to divide sensitive values into levels and assign each value to its level as done in [10] to be able to detect values that are semantically similar. On the other side, we treat the semantic similarity between sensitive categorical values according to the treated hierarchy without transforming the content of the hierarchy into semantic rules as done in [39].…”

“…e objective of this paper is to ensure privacy in a data set containing sensitive categorical attributes. However, the existing techniques using the principle of L-diversity for hierarchical data may suffer from various limitations and thus cannot ensure privacy carefully [10,11]. Among L-diversity for hierarchical data technique limitations, it does not take into account the semantic of categorical values.…”

“…e L-diversity technique splits the data set into several buckets including only distinct values while guaranteeing that there are at least L "well-represented" sensitive records [9]. However, the current anonymization approaches taking into account L-diversity for hierarchical data suffer from the similarity attack and cannot prevent the disclosure of data very well [10]. In spite of having buckets that contain only distinct values, these values may be semantically similar; thus, through a similarity attack, an adversary could easily violate the privacy of a certain individual by knowing his or her sensitive attributes' values.…”