A Deep Learning-Based Framework for Detecting and Classifying Malicious Network Activities in Complex Cybersecurity Environments

Abstract

The rapid growth of interconnected digital networks has increased the need for reliable and intelligent approaches to detect and classify malicious network activities in complex cybersecurity environments. This study developed a deep learning-based framework for detecting and classifying malicious network activities using network traffic data. A quantitative computational research design was employed using a cybersecurity network traffic dataset containing 10,000 network records collected from a publicly available benchmark environment. The dataset included normal and malicious network traffic and was prepared through data preprocessing, missing-value treatment, duplicate removal, feature transformation, and feature optimization. The dataset was divided into training, validation, and testing subsets, while relevant network traffic characteristics were selected as inputs for deep learning model development. DNN, CNN, RNN, and LSTM architectures were evaluated, and the proposed framework integrated preprocessing, feature optimization, malicious activity detection, multiclass attack classification, and performance evaluation. Python-based tools and libraries were used for data processing, model development, training, and evaluation, while accuracy, precision, recall, F1-score, confusion-matrix analysis, and computational performance were used to assess the framework without relying on statistical hypothesis testing. The findings demonstrated that the proposed framework was capable of distinguishing normal and malicious traffic and classifying multiple categories of malicious network activities, while comparative analysis demonstrated differences in the performance of the evaluated deep learning architectures. The study is significant because it provides an integrated computational approach for automated cybersecurity monitoring and can support the development of more adaptive and intelligent intrusion detection systems.