Project topics and materials
Get Complete Project Material File(s) Now! »
Background
Many organisations, including government departments and businesses, collect and analyse data related to individuals. The data is then used to assist in the planning and decision-making activities of those organisations. For example, governments may collect data about individuals by means of a census. The data may then be analysed and released in the form of statistical data to assist in the assessment of population trends and to guide the development of government policies (Zielinski, 2006, 2007a).
However, when personal data is collected, analysed, or released in the form of statistical data, it is necessary to protect the privacy of the individuals whose data is used. This is necessary not only to ensure ethical conduct, but also to respect different privacy and data protection laws. This need is especially evident in environments where the data is of a highly sensitive nature, as it is in, for example, the medical environment (Gostin & Turek-Brezina, 1995), commerce (Rauhofer, 2008; Paul, 2001), or in the context of eParticipation (Zielinski, 2007a).
Statistical data can be disseminated in three different ways. These include dynamically queryable databases, tabular data, and microdata (Hundepool et al., 2007; Domingo- Ferrer, Sebe, & Solanas, 2008; Willenborg & De Waal, 2001). However, in this research work, we focus on microdata, since it used as the basis from which all other statistical data outputs are derived. A microdata set is the « raw data » itself; it is a set of records, where each record contains information on the entities represented in the database.
To protect the privacy of the respondents whose data is released, it is not sufficient to de-identify the microdata set by removing explicit identifiers (e.g. an ID number) fromthe microdata set (Samarati, 2001; Skinner & Elliot, 2001). That is, a de-identified microdata set can still be manipulated and / or matched with external sources of data in an effort to re-identify individuals, or to disclose confidential data. Therefore, to protect privacy, a microdata set needs to be anonymised before it can be released.
Problem statement
To protect the privacy of the respondents in a microdata set, the microdata needs to be anonymised. As microdata is anonymised, data is removed (to some extent) from the identifying variables. As more data is removed from the identifying variables, it becomes increasingly difficult to infer sensitive data and to perform re-identification.
Therefore, as microdata is anonymised, the level of privacy in the microdata increases. However, removing data from the identifying variables also reduces the accuracy and / or completeness of the released microdata. Therefore, as microdata is anonymised, its level of information utility also decreases. Consequently, as we increase the level of privacy in a microdata set, the level of information utility decreases, and vice versa.
Ideally, we would like to release microdata that has high levels of privacy and information utility. However, the protection of privacy implies that we should hide and obscure data. On the other hand, releasing usable and useful data implies that we should provide data that is accurate, complete and precise (Zielinski, 2007a, 2007b). Clearly, a conflict between the needs of privacy and information utility exists. This conflict needs to be resolved before a microdata set can be released.
CHAPTER 1 INTRODUCTION
1.1 Background
1.2 Problem statement.
1.3 Research question, goal, and objectives of the study
1.4 Delimitation of scop
1.5 Methodology
1.6 Organisation of chapter
CHAPTER 2 DISSEMINATION OF STATISTICAL DATA .
2.1 Introduction
2.2 The statistical data life cycle
2.3 Ways in which statistical data is disseminated
2.4 Microdata
2.5 Conclusion
CHAPTER 3 PRIVACY IN THE CONTEXT OF INFORMATION SECURITY
3.1 Introduction
3.2 Privacy
3.2.1 Defining privacy
3.2.2 Why should information privacy be protected?.
3.3 How can privacy be compromised in microdata
3.4 Conclusion
CHAPTER 4 PRIVACY PROTECTION IN MICRODATA
4.1 Introduction.
4.2 Approaches for statistical disclosure control
4.2.1 Protection of data output through access contro
4.2.2 Protection of dynamically queryable databases
4.2.3 Protection of tabular data
4.2.4 Protection of microdata
4.3 Recoding
4.4 Microaggregation.
4.5 Conclusion
CHAPTER 5 THE CONFLICT BETWEEN PRIVACY AND INFORMATION UTILITY .
CHAPTER 6 HOW TO DETERMINE THE OPTIMUM LEVELS OF PRIVACY AND INFORMATION UTILITY
CHAPTER 7 HOW TO ANONYMISE MICRODATA TO ACHIEVE THE OPTIMUM LEVELS OF PRIVACY AND INFORMATION UTILITY
CHAPTER 8 CONCLUSION
