Dimensionality Reduction with DLMNN Technique for Handling Secure Medical Data in Healthcare-IoT Model

Security and maintaining the integrity of sensitive patient information have emerged as major obstacles for healthcare service applications in the wake of IoT’s meteoric rise in the healthcare industry. The diagnostic information included in medical photographs is of particular importance, and this research presents a hybrid security strategy to protect against such breaches. Chronic diseases such as Heart Disease (HD), cancer, and chronic respiratory sickness are the main causes of death worldwide. Those who have a heart attack suddenly have a very poor chance of surviving. As a result, this paper proposes a heart disease monitoring scheme, wherein a Deep Learning Modified Neural Network (DLMNN) centred on the Internet of Things is used to aid in the diagnosis and treatment of HD. There are three stages to implementing this strategy: First, there is encryption; second, a combination of information gain (IG) and principle component analysis (PCA) to reduce dimensionality; and third, a classification system. To begin, the patient is fitted with a wearable Internet of Things (IoT) sensor device, which simultaneously uploads its sensor data to a remote server. The sensor data is encrypted using a hybrid encryption scheme that combines the Advanced Encryption Standard (AES) and the Rivest, Shamir, and Adleman (RSA) algorithms. A DLMNN classifier is then used to classify the data after it has been decrypted. There are two types of classification outcomes. It provides information on the patient’s heart and notifies the doctor in case of abnormalities. Estimated study results demonstrate that DLMNN is an improvement over previously used algorithms for HD diagnosis. The proposed AES-RSA used to support safe data transfer also yields the highest level of security, i.e. 94.32%, and does so in the quickest time for encryption and decryption.