Abstract
Dense Wavelength Division Multiplexing (DWDM) networks are pivotal in modern optical communications, facilitating high data capacity by transmitting multiple wavelengths along a single optical fiber. Ensuring optimal network performance and reliability requires close monitoring of crucial signal characteristics such as signal-to-noise ratio (SNR), chromatic dispersion, nonlinear effects, and modulation types. However, traditional monitoring approaches often demand sophisticated and costly equipment, limiting their scalability. This paper proposes an innovative machine learning (ML)-based solution for enhanced signal monitoring and anomaly detection within DWDM networks. Utilizing K-means clustering for data categorization and feature extraction, alongside Support Vector Machine (SVM) for effective signal classification and anomaly detection, this approach provides scalable real-time monitoring while minimizing dependency on complex hardware. Additionally, this ML-driven framework allows for proactive network management, offering dynamic adaptability to network fluctuations. Our study emphasizes the promise of machine learning in optical network monitoring, paving the way for intelligent, self-optimizing systems that can support evolving communication demands and maintain superior service quality.