Heart Disease Detection in Cloud Platforms: A Privacy-Driven Approach using Exponential Distribution Optimized Hopfield Networks and Blockchain Security

In contemporary cloud health settings, guaranteeing heart disease prediction privacy and accuracy is of utmost importance in light of amplified exposure of personal patient data to risks. As a means to counteract the challenges faced, this research posits an innovative Blockchain-Assisted Privacy- Preserving Heart Disease Prediction Model employing Genetic- Encrypted Neural Network with Optimized Hopfield Framework (BAPP-HDP-GONet). First, clinical data are collected from UCI Repository Statlog and Cleveland Heart Disease databases, and critical feature representations are obtained from a Modified ResNet-152 Network to achieve high-dimensional medical patterns. In order to make the extracted features trustworthy and free from tampering, a Blockchain-based Fair Proof-of-Reputation (PoR) mechanism is implemented, which authenticates and verifies only credible data sources for prediction. Also, a Lightweight Genetic-Based Encryption (LGAE) method is applied to encrypt sensitive feature data prior to cloud storage. The framework uses a Discrete Hopfield Neural Network (DHNN) for preliminary heart disease classification, but as common DHNN suffers from limitations of parameter optimization, an Exponential Distribution Optimizer (EDO) is introduced to optimize DHNN's weights and thresholds in order to provide maximum diagnostic accuracy. The suggested BAPP-HDP-GONet model attains 99.99% greater accuracy and 99.98% greater precision compared to current methods, thereby providing an efficient privacy-preserving and high-performance solution for heart disease prediction in cloud-based health care systems.