Institute of Engineering, Suranaree University of Technology, Nakhon Ratchasima, Thailand
Email：d6700775@g.sut.ac.th (P.A.); nuntawut@sut.ac.th (N.K.); parin.s@sut.ac.th (P.S.)
*Corresponding author
Manuscript received November 15, 2025; accepted December 23, 2025; published January 26, 2026
Abstract—Transient testing of the Erbium-Doped Fiber Amplifier (EDFA) is crucial for assessing the transient response to abrupt channel add/drop occurrences in Dense Wavelength Division Multiplexing (DWDM) systems. Nevertheless, anomalous transient signals, frequently resulting from inadequacies in the testing apparatus or photodetector saturation, may result in erroneous computations of critical parameters, including overshoot, undershoot, gain offset, and settling time. This study presents an artificial intelligence-driven method for the automatic classification of normal and pathological transient signals in EDFA testing. The methodology entails preprocessing transient signals via uniform downsampling, implementing feature extraction techniques such as Classical Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), and Partial Least Squares–Discriminant Analysis (PLS-DA), and subsequently training and assessing five machine learning models (Naïve Bayes, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Logistic Regression, and Multi-Layer Perceptron (MLP)) for performance evaluation. Experimental findings demonstrate that the number of features and the extraction techniques substantially affect classification accuracy, Area Under the Curve (AUC), precision, recall, F1-score, and processing duration. The optimal performance was attained using 1000 features produced through uniform downsampling, integrated with PLS-DA and categorized using SVM, resulting in an AUC of 0.9977. The findings illustrate a dependable and effective AI-driven method for automated classification of transient signals, augmenting the validation of EDFA transient testing and potentially enhancing signal data analysis in the assessment of optical fiber communication systems.
Keywords—artificial intelligence, Erbium-Doped Fiber Amplifier (EDFA), feature extraction, machine learning, transient signal, uniform downsampling
[PDF]
Cite: Peerasut Anmahapong, Nuntawut Kaoungku, and Parin Sornlertlamvanich, "An AI-Driven Methodology for Classifying Transient Signals in the Evaluation of Optical Fiber Amplifiers," International Journal of Future Computer and Communication, vol. 15, no. 1, pp. 1-10, 2026.
Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited
(CC BY 4.0)
