A Real-Time Deep Neural Network–Based Architecture for Active Network Attack Identification and Classification Models

In the context of computer science and technology, a group of linked devices, or nodes, that exchange data, resources, or services with one another is referred to as a network. An active network attack refers to a malicious activity in which an attacker deliberately attempts to disrupt, manipulate, or gain unauthorized access to a computer network or its resources. In today’s digital context, network security is of vital importance as cyber threats continue to evolve in terms of sophistication and frequency. Active network attacks pose significant challenges to traditional detection methods, necessitating the exploration of advanced techniques such as deep learning.  The proposed methodology involves the development of a model based on deep learning that was learned using a dataset comprising diverse network traffic data which is Network Security Laboratory-Knowledge Discovery in Databases (NSL-KDD). This study utilizes a comprehensive preprocessing pipeline, including data cleaning, feature selection for categorical variables and standardization of numerical features to prepare the dataset for modeling. To extract the pertinent information, preprocessing approaches are used. Metrics like as accuracy, precision, recall, F1-Score, and confusion matrix are used to evaluate performance as a result from deep learning models Deep Neural Network (DNN), Convolution Neural Networks (CNN), Long Short Term Memory(LSTM), Bi-Long Short Term Memory (Bi-LSTM) and Gated Recurrent Units (GRU) experiments done, Bi-LSTM model scored the best result of 99.15% and 99.12% accuracy for binary and multi classification, respectively.