Cyber Security Intrusion Detection Using Deep Learning Approaches, Datasets, Bot-IOT Dataset

Abstract

Cyber Security is a crucial point of the current world; it is used to analyze, defend, and detect network intrusion systems. An intrusion detection system has been designed using Deep learning techniques, which helps the network user to detect malicious intentions. The dataset plays a crucial part in intrusion detection. As a result, we describe various well-known cyber datasets. Mainly we have analysed the IoT traffic-based dataset with some other datasets. We have also analyzed deep learning DL models, including Feed forward deep neural network (FDNN), deep auto-encoder, De-noising auto-encoder, deep migration, stacked de-noising auto-encoders, Replicator Neural networks, and Self-Taught Learning. We observe the effectiveness of models individually in two different types (multiclass and binary) through real-world traffic datasets, such as Bot-IoT dataset. Moreover, we evaluate the effectiveness of various methods based on the most critical key performance indicators, namely correctness, rate of false alarms and detection rate.