A Novel Framework of Anomaly-Based Network Intrusion Detection using Hybrid CNN, Bi-LSTM Deep Learning Techniques

Abstract

A Novel Framework of Anomaly-based Network Intrusion Detection system using hybrid CNN,Bi-LSTM Deep learning techniques with the aim of anomaly detection, In recent years, deep learning (DL) has become increasingly important in the field of cyber security. Deep learning Algorithms efficient to detect vulnerabilities in network traffic.Objective are based on literature survey provides the various anomaly based techniques Such as NIDS,SIDS, researches are presented. [proposed CNN based BLISTM model]stand out, providing a solid basis for understanding the context of the investigation and verified results with slandered existing systems and studies.The methodology adopted for this research comprises [a hybrid CNN-based BLISTM model with Adam optimizer was used] and [the well-known NSL KDD data set was used to validate the proposed modelThe results obtained revealed that efficacy of the suggested CNN-Bi-LSTM IDS has been assessed for the NSL-KDD dataset. Imbalanced data fed into CNN-Bi-LSTM accuracy achieved 98 % recall 98% and precision 99 %, F1-score98 %, After balanced data and hyper parameter tuning of CNN-Bi-LSTM classifier, Exceptional accuracy was demonstrated by the binary classification results, which included a 99.12% accuracy for the NSL KDD dataset with the precision of 99.0%, recall of 99.26%, and F1-score of 98.11%.. Conclusions  are  a novel frame work enhanced accuracy in detection of anomalies in network traffic in the field of[Network  security]. These implications could encompass list impacted are such as Medical and Banking, Ecommerce sites.This study contributes to the literature by hybrid CNN based BLISTM generates efficacious results. The relevance and value of this research are evidenced by comparison of generated results with existing literature results. In future  work it would applied for various different datasets .