No abstract

No abstract

IoT forensics where security and privacy are the key concern as the data the majorly hold personal information. So how to work on the vulnerabilities available from the IoT environment and classify them to get the best results to perform the forensics is covered in the paper. In IoT forensics botnet dataset analysed using deep learning classification to get the understanding that how deep learning can be used effectively for forensic analysis. So, research work provides advanced digital forensics methods i.e., collection of evidence and analysis of dataset for IoT forensics implementation. Since a decade ago, we are seeing a reality where hacking into a client's PC utilizing small bots or blocking a gathering of interconnected gadgets is not any more unthinkable. These little bots are called botnets (e.g., Mirai, Torii and so on.), which are a gathering of deadly codes that can obstruct the whole security. As Internet of Things (IoT) is developing quickly, the interconnected gadgets are helpless to penetrate as one influenced gadget can crumple the entire system. As Internet of Things (IoT) is developing quickly, the interconnected gadgets are defenseless to break as one influenced gadget can hamper the entire system. The security danger stays as botnet assaults increment their essence to the interconnected gadgets. In this work, we are proposing a novel correlation between AI (SVM and KNN) and profound learning draws near (Neural system) to discover which approach creates better outcome while learning the assault designs. Research explores the IoT forensics analysis. In IoT forensics models were applied on a composite information storehouse which was made by consolidating the outcomes found from the examination we did on Torii botnet test, with the CTU-13 dataset of botnet assaults on IoT environment.

As Ransomware encrypts user files to prevent access to infected systems its harmful impacts must be quickly identified and remedied. It can be challenging to identify the metrics and parameters to check, especially when using unknown ransomware variants in tests. The proposed work uses machine learning techniques to create a general model that can be used to detect the variations of ransomware families while observing the characteristics of malware. However, early detection is impeded by a dearth of data during the initial phases of an attack, which results in low detection accuracy and a high proportion of false alarms.To overcome these restrictions, our research suggests a revolutionary technique, in machine learning techniques we have proposedRandomClassifier with SMOTE optimizer based on the results received from LazyPredictAutoML and then deep learning algorithm ANN using Root Mean Square Propagation (adam) has been implemented to get the hidden patterns which were not accessible in machine learning approach. Further study focused on improving CNN's performance over RMSProp& Adam, which maintains per-parameter learning rates that are adjusted based on the average of most recent weight gradient magnitudes, using the Adam optimizer. The best option for internet and non-stationary issues is CNN with Adam (e.g. noisy). As gradients grow sparser toward the end of optimization, Adam somewhat surpasses RMSprop. Adam uses CNN and uses the average of the second moments of the gradients (the uncentered variance). The proposed model reported 5.14ms of prediction time and 99.18% accuracy.

The untraceable part of the Deep Web, also known as the Dark Web, is one of the most used “secretive spaces” to execute all sorts of illegal and criminal activities by terrorists, cybercriminals, spies, and offenders. Identifying actions, products, and offenders on the Dark Web is challenging due to its size, intractability, and anonymity. Therefore, it is crucial to intelligently enforce tools and techniques capable of identifying the activities of the Dark Web to assist law enforcement agencies as a support system. Therefore, this study proposes four deep learning architectures (RNN, CNN, LSTM, and Transformer)-based classification models using the pre-trained word embedding representations to identify illicit activities related to cybercrimes on Dark Web forums. We used the Agora dataset derived from the DarkNet market archive, which lists 109 activities by category. The listings in the dataset are vaguely described, and several data points are untagged, which rules out the automatic labeling of category items as target classes. Hence, to overcome this constraint, we applied a meticulously designed human annotation scheme to annotate the data, taking into account all the attributes to infer the context. In this research, we conducted comprehensive evaluations to assess the performance of our proposed approach. Our proposed BERT-based classification model achieved an accuracy score of 96%. Given the unbalancedness of the experimental data, our results indicate the advantage of our tailored data preprocessing strategies and validate our annotation scheme. Thus, in real-world scenarios, our work can be used to analyze Dark Web forums and identify cybercrimes by law enforcement agencies and can pave the path to develop sophisticated systems as per the requirements.