A Secure Blockchain‐Based Health Prediction Framework Using Mantaray Optimization

ABSTRACT One of the leading causes of various critical diseases for both men and women worldwide is obesity. This study uses a publicly available dataset to predict a person's weight category—normal, overweight, underweight, or obese—by examining lifestyle factors and daily activities. There is a need for new and sophisticated secure algorithms for quick processing and effective event detection. This paper subjects the dataset to a thorough feature selection and engineering process to identify the key factors influencing weight status. For obesity prediction, two novel machine learning‐based models are proposed: the neural network boost (N_XB) and the Mantaray optimization (MO‐NXB), which learn from the data stored in the blockchain. The blockchain serves as a safe storage environment for patient information and a genuine source for learning data because its data is impenetrable. The accuracy, precision, recall, and F1‐score of the suggested N_XB and MO‐NXB algorithms are evaluated by other algorithms. According to our investigation, the greatest accuracy of N_XB and MO‐NXB is 0.94 and 0.985, which is approximately 10% more than traditional models. Additionally, the latency of blockchain‐based storage, when contrasted with centralized storage, are 25.03% less than those of centralized storage.