A Novel Approach to Secure Data Sharing

Abstract

Nowadays data sharing utilization is in research and development and many sectors like health, defense,
machine learning, etc. Many institution or organizations have their policies to share data without loss of confidentiality
of their data. Sharing of information across databases maintained by different organizations leads to an exchange of
personal information about an individual. In practice, various statutory regulations and policies prevent the disclosure
of such identifiers. So, to prevent the confidentiality of the data, it should share securely. Multiple techniques are
implemented to protect the data from attack while sharing. K-anonymization, de-identification of personnel identifiers,
noise addition for sharing data have become a topic of recent interest in many organizations due to their versatility and
ability to share the data securely. A novel approach for data sharing has been worked out by combining the kanonymization and addition of noise in the dataset. K-anonymization is applied on personnel identifiers and addition of
random noise is applied to numerical data to secure the data for sharing. An application was developed to demonstrate
the proposed concept.