Hybrid Machine Learning Models for Real-Time Anomaly Detection in Complex Deployment Environments

Abstract

Anomaly detection plays a crucial role in maintaining the integrity and security of real-time systems in diverse application areas, including cybersecurity, predictive maintenance, healthcare, and IoT networks. Traditional machine learning models, such as Deep Neural Networks (DNNs) and Support Vector Machines (SVMs), often struggle with high-dimensional, noisy data and the need for real-time processing, making them less effective in dynamic deployment environments. This paper presents a hybrid machine learning model that integrates Ensemble Decision Trees and K-Nearest Neighbors (KNN) to address these challenges. The Decision Tree model is employed for initial classification based on global data patterns, while KNN is used to refine anomaly detection by focusing on local relationships between data points. The proposed model is evaluated on the KDD Cup 1999 dataset, a widely-used benchmark for anomaly detection in network traffic, and achieves a high accuracy of 98.69%. The hybrid approach demonstrates both high detection accuracy and computational efficiency, making it suitable for real-time anomaly detection in complex environments. The paper also discusses how the integration of feature selection and real-time adaptation further enhances the model's performance, ensuring its applicability across various domains, including IoT systems, healthcare monitoring, cybersecurity, and predictive maintenance.