Design of Deep Neural Network based Anomaly Detection System
Rounak Gupta, Zha Chanakya Kumar, Vijay Narendranath Patil · 2022 IEEE 7th International conference for Convergence in Technology (I2CT) · 2022
A novel anomaly detection-based NIDS is main Demand in the computer networking security for discriminating Malicious software attack at the early stage. It monitors and Analyzes network traffics, checking abnormal behaviors or attack Signatures. Unknown Attacks need high detection rates and accuracy in network intrusion detection models. With software defined networking, deep neural networks (DNN) can identify anomalies. The dropout method prevents overfitting DNN models. About six characteristics have data. The NSL-KDD dataset was used to fit and assess the flow. These data attributes allow the model to be excellent generative and perform well on intrusion recognition using a fraction of the data. Lessons learned from cross-entropy loss function with SoftMax output layer Five class labels are covered by two distributions, one normal and two assaults (Dos, R2L, U2L and Probe). Accuracy is a model performance statistic. A 92.65% accuracy rate is promising.