Network Attack Detection Using Long-Short Term Memory (LSTM)

Network Attack Detection Using Long-Short Term Memory (LSTM)

Jyoti Chandel, Syad Abdul Safi, Vagesh Kumar, Hemant Kashyap, Isha Paul, Dhawal Tiwari

Chandigarh University

Punjab, India

Abstract-

In an age of sophisticated cyber threats, the need for a good    network intrusion detection system is crucial. This research    aims to push the boundaries of cyber attack detection by leveraging the power of short term memory (LSTM) neural networks in the context of deep learning. The focus here is on different types of classification that solve the difficult task of classifying different types of     cyber attacks . This approach goes beyond traditional binary classification and can be adapted to the nuances of cyberattacks. LSTM- based models that use realtime predictions on network traffic  data show significant improvements in accuracy and  adaptability. We conducted extensive testing to finetune the hyperparameters, adjust the data before the procedure, and improve the model design. The results demonstrate the effectiveness of this approach in classifying cyber attacks and minimizing the possibility of vulnerability.

This research contributes to the continuous improvement of network intrusion detection by combining the power of LSTM based deep learning models with integrated processing. It  shows that different types of classification are possible using LSTM for cyber attack detection and paves the way for  further developments in the field of cyber security.

Keywords:Multivariate classification, network attack  detection, LSTM, deep learning, network penetration  detection, network security, time dependency, accuracy,  adaptability, data sheet preprocessing technology, model architecture, fake quality.