Big Data Analytics Cyber Forensic Framework Using Enhanced Sunflower Jaya Optimization- Based Deep Stacked Autoencoder

The skills of forensic analysts are being threatened by the growing volume of information on the Internet of Things (IoT) environment. However, efficacy of the current network forensics frameworks is impacted by technical challenges such as anti-forensics, diverse traffic formats, steganography, real-time investigations and encrypted data. An efficient method known as Enhanced Sunflower Jaya Optimization with Deep Stacked Autoencoder Algorithm (ESFJO) is suggested for network forensics frameworks in order to address these problems. To overcome the difficulties in the realm of network forensics, the Sunflower optimized FE model is coupled by Jaya optimized control parameters. By reducing mistakes in massive data analysis models, the suggested ESFJO-based Deep Stacked Autoencoder enhances the network forensics framework by taking use of pollination behaviours and other special characteristics. As a result, the optimum solution identified by computing the direction vector is the one with the lowest error value. An unrestrained benchmark function, which yields a fitness function that determines the optimal candidate solution, is used to illustrate the suggested model. The ESFJO-based deep stacked autoencoder performs admirably in terms of accuracy, sensitivity, and specificity; on UNSWNB15, it received scores of 0.9183, 0.9765, and 0.8909.