Abstract：
Numerous studies have been conducted in an attempt to preserve cloud privacy, yet the majority of cutting-edge solutions fall short when it comes to handling sensitive data. This research proposes a \"privacy preservation model in the cloud environment\". The four stages of recommended security preservation methodology are \"identification of sensitive data, generation of an optimal tuned key, suggested data sanitization, and data restoration\". Initially, owner\'s data enters the Sensitive data identification process. The sensitive information in the input (owner\'s data) is identified via Augmented Dynamic Itemset Counting (ADIC) based Associative Rule Mining Model. Subsequently, the identified sensitive data are sanitized via the newly created tuned key. The generated tuned key is formulated with new fourfold objective-hybrid optimization approach-based deep learning approach. The optimally tuned key is generated with LSTM on the basis of fourfold objectives and the new hybrid MUAOA. The created keys, as well as generated sensitive rules, are fed into the deep learning model. The MUAOA technique is a conceptual blend of standard AOA and CMBO, respectively. As a result, unauthorized people will be unable to access information. Finally, comparative evaluation is undergone and proposed LSTM+MUAOA has achieved higher values on privacy about 5.21 compared to other existing models.
摘要：
为了保护云隐私，已经进行了许多研究，然而，在处理敏感数据方面，大多数尖端解决方案都不够。本研究提出了“云环境下的隐私保护模型”。建议的安全保存方法的四个阶段是“识别敏感数据，生成最佳调谐密钥，建议的数据清理，和数据恢复\"。最初，所有者的数据进入敏感数据识别过程。输入中的敏感信息(所有者数据)通过基于关联规则挖掘模型的增强动态项集计数(ADIC)来识别。随后，识别的敏感数据通过新创建的调整密钥进行清理。生成的调整密钥是用新的基于四重目标混合优化方法的深度学习方法制定的。最佳调谐密钥是在四重目标和新的混合MUAOA的基础上用LSTM生成的。创建的密钥，以及生成的敏感规则，被馈送到深度学习模型中。MUAOA技术是标准AOA和CMBO的概念融合，分别。因此,未经授权的人将无法访问信息。最后,进行了比较评估，与其他现有模型相比，提出的LSTM+MUAOA在隐私方面取得了更高的价值约5.21。
