Artificial Neural Network based Data sensitive Security Model for Healthcare Data

Abstract

Using artificial neural networks, we propose data-sensitive security architecture for healthcare data sets of unprecedented size and complexity. Modern medical gadgets and the digitization of healthcare data have added new layers of complexity to data analytics. A large amount of data is generated by this procedure, which is then usefully analyzed and categorized to provide useful insights. Using this information and the results of the study, medical diagnoses and prognoses can be made. Many countries' healthcare laws specifically require the protection and storage of patient records. Therefore, the healthcare industry places a premium on protecting the confidentiality of this vast trove of information. The same level of security applies anywhere else, too. Conversely, the factors' relative relevance determines the data's sensitivity, which is not constant. Therefore, it has been shown that employing a solitary security architecture for all data is redundant and inefficient. The adaptive intelligent security system proposed in this study will tailor protection to the specifics of the stored data. Big Data, Security, and Machine Learning, the three primary pillars of the proposed architecture, are mapped to accomplish varying degrees of data confidentiality. Due to the unpredictable and unstructured nature of medical data, an artificial neural network is trained using an electronic medical record and a sensitivity level to produce a novel security model that offers the best security in real time. The ANN is trained with attributes from data sets that include patient information, medical history, insurance information, and so on, and accurate categorization is achieved. The proposed method is tested experimentally to determine its efficacy.