The reliability of the digital forensic system mainly lies in its ability to detect the attack on time and minimize the security risk associated with the Internet of Things (IoT) device. The existing machine learning architectures mainly suffered from different complexities while processing the attack features such as increased processing time, high manual intervention, complexity to handle large datasets, and low prediction accuracy. To overcome these issues, this article proposes a novel deep learning architecture for data forensics. The low generalization capability of the neural network is handled via the Hadoop clustering and convolution‐based ADAM (CADAM) optimizer approach for improving the optimization capability of the model. In this way, the privacy of the model is improved. The efficiency of the proposed CADAM optimized deep neural network architecture is evaluated by using the UNSW‐NB15 and Bot‐IoT datasets. The proposed model offers improved performance in terms of runtime, precision, sensitivity, specificity, latency, cost, ROCAUC curve, and anomalous traffic when compared with the existing techniques. From the experiments carried out, we can conclude that the proposed framework revealed higher performance to discover and prevent cyber‐attacks against other conventional approaches.
The skills of forensic analysts are at risk to process the increasing data in the Internet of Things‐based environment platforms. However, the technical issues like anti‐forensics, variety of traffic formats, steganography or encrypted data, and real‐time live investigation degrades the performance of the cyber forensic framework. Therefore, an effective method named Sunflower Jaya Optimization‐based Deep stacked autoencoder (SFJO‐based Deep stacked autoencoder) is proposed to perform the cyber forensic framework. The finite element model of Sunflower optimization is integrated with the control parameters of Jaya optimization to solve the issues in the cyber forensic framework. The proposed SFJO‐based Deep stacked autoencoder uses the pollination and the peculiar behaviors to enable the cyber forensic framework based on the error value in the big data analytics model. Accordingly, the solution with the minimal value of error is accepted as the best optimal solution by computing the orientation vector. However, the proposed model is illustrated based on the unconstrained benchmark function, which in turn results in the fitness function to reveal the best candidate solution. The proposed SFJO‐based Deep stacked autoencoder attained better performance using metrics like precision, sensitivity, and specificity with the values of 0.9053, 0.8865, and 0.8839 using dataset‐1.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.