Machine Learning Based Framework for Maintaining Privacy of Healthcare Data
The Adoption of Artificial Intelligence (AI), Machine Learning (ML), Internet of Things (IoT), cloud services, web-based software systems, and other wireless sensor devices in the healthcare infrastructure have led to phenomenal improvements and benefits in the healthcare sector. Digital healthcare has ensured early diagnosis of the diseases, greater accessibility, and mass outreach in terms of treatment. Despite this unprecedented success, the privacy and confidentiality of the healthcare data have become a major concern for all the stakeholders. Data breach reports reveal that the healthcare data industry is one of the key targets of cyber invaders. In fact the last few years have registered an unprecedented rise in healthcare data breaches. Hacking incidents and privilege abuse are the most common threats and have exposed sensitive and protected health data. Experts and researchers are working on various techniques, tools, and methods to address the security issues related to healthcare data. In this article, the main focus is on evaluating the impact of research studies done in the context of healthcare data breach reports to identify the contemporary privacy and confidentiality issues of sensitive healthcare data. Analysis of the research studies depicts that there is a need for proactive security mechanisms that will help the healthcare organizations to identify abnormal user behavior while accessing healthcare data. Moreover, studies also suggest that ML techniques would be highly effective in securing the privacy and confidentiality of the healthcare data. Working further on this premise, the present study also proposes a conceptual framework that will secure the privacy and confidentiality of healthcare data proactively. The proposed framework is based on ML techniques to detect deviated user access against Electronic Health Records. Further, fuzzy-based Analytical Network Process (ANP), a multi-criteria decision-making approach, is used to assess the accuracy of the supervised and unsupervised ML approaches for achieving a dynamic digital healthcare data security environment.