Optimized Nonlocal Kernel Network‐Based Keyword Search for Comprehensive Detection and Prevention of Forward Security Threats in Encrypted Cloud Data Systems

ABSTRACT Keyword search over encrypted information allows workers to quickly locate the most relevant outcomes; cloud computing researchers have researched keyword search over encrypted data extensively. The effective ciphertext search and dynamic updates by forward security cannot be achieved at the same time by currently available ranked multikeyword search systems. In this manuscript, an Optimized Nonlocal Kernel Network‐Based Keyword Search for Comprehensive Detection and Prevention of Forward Security Threats in Encrypted Cloud Data Systems (ONKN‐KS‐FST‐ECD) is proposed. The ONKN‐KS‐FST‐ECD model uses the Nonlocal Kernel Network (NKN) approach to achieve forward security by reducing search complexity while maintaining search accuracy and by preventing cloud servers from using earlier tokens to make search queries over newly uploaded files. The Circulatory System‐Based Optimization Algorithm (CSBOA) is used to improve the weight parameter of the NKN model to increase the search accuracy. The proposed ONKN‐KS‐FST‐ECD method attains 24.28%, 28.22%, and 29.27% higher search accuracy and lower file updating time of range 14.76%, 16.82%, and 12.47% compared to existing techniques such as ranked keyword search over encrypted cloud data over machine learning approach (RKS‐ECD‐ML), a system for privacy‐preserving and effectiveness search over encrypted social graphs (PPKSS‐ESD‐CC), and multiclient secure and efficiency dpf‐dependent keyword search for cloud storage (DDPS‐SCDSC) respectively.