DNNBoT: Deep Neural Network-Based Botnet Detection and Classification

The evolution and expansion of IoT devices reduced human efforts, increased resource utilization, saved time; however, create significant challenges such as lack security privacy, making them more vulnerable to IoT-based botnet attacks. There is a need develop efficient faster models which can work in real-time with efficiency stability. present investigation developed two novels, Deep Neural Network (DNN) models, DNNBoT1 DNNBoT2, detect classify well-known attacks Mirai BASHLITE from nine compromised industrial-grade devices. utilization PCA was made feature extraction improve effectual accurate Botnet classification environments. were designed based on rigorous hyperparameters tuning GridsearchCV. Early stopping utilized avoid the effects overfitting underfitting for both DNN models. in-depth assessment evaluation demonstrated that accuracy are some best-performed novelty investigation, bridge gaps by using real dataset high significantly lower false alarm rate. results evaluated earlier studies deemed at detecting dataset.