Hybrid machine learning model for network traffic anomaly detection using time-series forecasting

This research study employs a hybrid model capable of detecting anomalies in network infrastructure to enhance cyber security. The proposed model analyzes anomalies and classifies cyber-attacks by combining ARIMA for time-series forecasting with advanced AI (Artificial Intelligence) models like autoencoders and Isolation Forests. ARIMA generates residuals based on deviations from predictions by capturing standard traffic patterns, which is then classified by AI models. The proposed approach leverages ARIMA’s temporal dependency handling and robust AI classification along with addressing limitations in