Study on preserving privacy and minimizing information loss in data publishing

Abstract

The development of information technology and bulk electronic information owned by governments, corporations and individuals have resulted in greater need for data sharing. Huge amount of information sharing between people all over the world is increasing due to digital technology. Several organizations frequently combine with other entities for performing data distribution. In Healthcare domain, there is an emerging need for distribution of data which is constrained by the presence of personal information in database. The information shared between health organizations such as hospitals supports the use of Health information for enhancing the privacy. Privacy Preserving Data Publishing is the method to ensure that the presented data remains useful and protected. However, the PPDP methods remove the recognizing attributes like name and id but other attributes like gender, age, zipcode are collaborated to identify the individual and the confidential information such as disease and salary. Several methods are used to provide privacy for maintaining the database and reduce the attacks during data distribution. Privacy preservation can be used to identify abnormal behaviour in data sharing. Many privacy-preserving data publishing techniques were developed but failed to consider the medical dataset because of its complexity and domain specific nature. Various organizations release information about persons in public for resource sharing. But maintaining the confidentiality of individual information is a difficult task with various data releases from multiple organizations where coordinating before data publication. Several research works namely k-anonymity, l-diversity model were developed but avoidance of the above said composition attack is a challenging issue during data publication. newline