Integrating K-Anonymity with Meta-Heuristic Algorithms for Enhanced Data Privacy and Security

The Internet of Things (IoT) is a rapidly evolving emergent technology used in the digitalization of Electronic Health Records (EHR). IoT applications are used to gather patients and data holder information before publishing it. However, the information gathered by IoT-based devices is susceptible to data leakage and poses a possible risk to privacy. To eliminate the recognition of individual records in EHR, privacy protection techniques must be implemented. Recently, a new model was developed that has less information loss than previous models; this is a kind of non-homogeneous generalization. In this work, we provide a different anonymization strategy with three classifiers, including Random Forest (RF), Support Vector Machine (SVM), and Logistic Regression (LR) which further minimizes information loss by employing black widow optimization methods. Sensitive information about an individual that is published and cannot be differentiated from at least k1 other individuals is referred to as k-anonymity with regard to the privacy of sensitive data. Clustering methods accurately achieve k-Anonymity. The machine learning classifier obtains 0.5324 accuracy for the diabetes dataset and 0.5614 accuracy for the heart disease dataset. Notably, this study addresses the critical challenges of data leakage and privacy risks inherent in IoT applications for EHRs. By developing an effective anonymisation strategy, the aim is to enhance data protection while maintaining the utility of health information.