Clustering based anonymization methods using utility constraints to preserve privacy

Abstract

newline Advances in data collection techniques and the need for automation prompted in the proliferation of a large amount of data. This exponential increase in the collection of personal information poses a serious threat to privacy. The problem was further enhanced with the advancement of technologies for data storage, data mining, machine learning, social networking and cloud computing. Privacy is a vital right of every human being and needs to be preserved. Most nations have both general policies on preserving privacy and specific legislation to control access to and use of data. Privacy preserving data publishing is the ability to regulate the dissemination of personal information. Publishing of raw data results in identity disclosure with linkage attacks. Statistical techniques are employed to overcome linkage attacks. k-anonymity model reduces the data across a set of attributes to a set of classes. In a k-anonymized dataset, each record is indistinguishable from at least k-1 other instances. But there is a tradeoff between data privacy and utility. Hence, a privacy preserving technique should ensure a balance of utility and privacy. The significant purpose behind the thesis is to design privacy preserving methods to efficiently preserve data utility while providing protection to the personal information.