Advanced Cryptographic and Clustering Techniques for Privacy and Anomaly Detection in Heterogeneous Cloud Environments

The study delves at machine learning methods that can spot out-of-the-ordinary occurrences in real life. The development of a secure anomaly detection system that enhances deployment-ready classification accuracy and execution speed is one goal of this endeavor. According to studies, data privacy and classification accuracy may both be enhanced with the right cryptography approach. Two components of privacy protection in commercial cloud systems are processing of information security and integrity. While processing, data processing security safeguards user data, and data integrity guarantees that data syntax and semantics remain unaffected while being sent. Hackers may take use of user data by breaking into key pairs using existing cryptography methods. There is a lack of consideration for noisy data in current anomaly detection systems. In response to the difficulties encountered in scientific research, we presented a low-computing-cost method for protecting user privacy while simultaneously detecting anomalies in cloud data. An efficient privacy-protecting cloud-based anomaly detection approach is introduced. For data collected by sensors on a massive scale, the method enhances cryptographic techniques. Data encryption is performed in the cloud on both individual and group servers. A more secure and private encryption method called RSA based on LPGWO (Local Pollination Grey Wolf Optimizer) is used for the encryption. To efficiently classify large datasets, the proposed method partitions them. Collaborative public servers detect anomalies using the NCM (Neutrosophic C-Means) method.