A novel Jarratt butterfly Ebola optimization-based attentional random forest for data anonymization in cloud environment

• Published in: The Journal of supercomputing 2024-03, Vol.80 (5), p.5950-5978
• Main Authors: Bushra, S. Nikkath, Subramanian, Nalini, Shobana, G., Radhika, S.
• Format: Article
• Language: English
• Subjects: Cloud computing; Compilers; Complexity; Computer Science; Cybersecurity; Data integrity; Data loss; Datasets; Decision trees; Ebola virus; Efficiency; Interpreters; Leakage; Optimization; Privacy; Processor Architectures; Programming Languages; Search algorithms
• ISSN: 0920-8542; 1573-0484
• DOI: 10.1007/s11227-023-05606-4

Privacy and security are the most essential barriers preventing widespread public cloud adaptation. The data that is distributed within the cloud requires privacy protection which can be achieved through the implementation of a data anonymization process. However, there is a lack of privacy for the user data. Even though the differential privacy model minimizes the efficiency, it also leads to generate data loss and privacy is not enhanced. Data securing in a cloud environment is a complex task. The data that are distributed within the cloud requires privacy protection which can be achieved through the implementation of a data anonymization process. So in order to improve the efficiency the novel hybrid ABRF-Stacked Bi-GRU is proposed to secure the medical data. This proposed method diminished the leakage of data and also effectively improve security and privacy. Therefore, in this paper, the privacy-preserving hybrid attention-based random forest (ABRF) with stacked bidirectional gated recurrent unit (stacked Bi-GRU) model named hybrid ABRF-Stacked Bi-GRU is proposed in the cloud environment. This proposed model is used to preserve medical data in the cloud from private data leakage. The attention weights along with the trainable parameters are assigned by the ABRF model to the decision trees. The privacy leakage is reduced by combining Jarratt butterfly optimization algorithm with the Ebola optimization search algorithm. However, in various existing methods, the differential privacy model is applied but it does not minimize the efficiency. In such cases it leads to generate data loss and privacy is not enhanced. Data securing in a cloud environment is a complex task. So in order to improve the efficiency the novel hybrid ABRF-Stacked Bi-GRU is proposed to secure the medical data. This proposed method diminished the leakage of data and also effectively improve security and privacy. The ensemble of the ABRF and stacked Bi-GRU model enhances the security of data in a cloud environment. The proposed model is evaluated using Diabetes UCI Dataset and Obesity UCI ML Dataset. The obtained outcomes compared with existing methods showed that the proposed method achieves a high anonymization degree with less Information Loss and takes less time for computation and minimizes the complexity during computation.