Abstract:
Privacy preservation is one of the key roles from security perspective in any data security environment. The purpose of this thesis is to provide privacy aspect of security in such a way that it provides strong Patient Anonymity Level, Anonymized Data Searching and successful Correlation of PHR for Medical Research in a single framework. Moreover, a novel solution for data de-identification is introduced (i.e., L-Diversity along with K-Anonymity) for anonymized data searching because previous method of using K-Anonymity (alone) is vulnerable to two type of attacks (homogeneity attack and background knowledge attack). Furthermore, it is experimentally proved in this research that using K-Anonymity alone can risk the disclosure of a huge number of medical records compared to L-Diversity along with K-Anonymity. The percentage risk analysis results are verified as well by using another dataset. Lastly, the experimental setup meets the requirements of HIPAA Privacy Rule as the attributes used in this research are specified by HIPAA as identifying attributes. These identifying attributes are totally suppressed (hidden) as per HIPAA requirement.
