Enhancing fault detection and classification in optical fiber networks with deep learning algorithms

What is it about?

The study attempts to assess the efficacy of deep learning (DL) architectures, employing Dense, Bidirectional-Long Short Term Memory (BiLSTM-Dense), LSTM-Dense,(Convolutional Neural Network) CNN-LSTM, and CNN-BiLSTM, in terms of performance. The findings indicate that Hist Gradient Boosting, Random Forest, and Extra Trees demonstrate exceptional precision in detecting and categorizing faults. CNN-BiLSTM and CNN-LSTM are considered the most effective DL models, as they exhibit exceptional performance by attaining high accuracy, precision, and recall levels. This study offers valuable contributions by shedding light on the appropriateness of various models for detecting and classifying faults in optical fiber.

Featured Image

Why is it important?

Detecting and classifying faults in optical fiber networks is essential for maintaining their performance and uninterrupted service, as they are vital communication infrastructures. CNN-BiLSTM and CNN-LSTM are considered the most effective DL models, as they exhibit exceptional performance by attaining high accuracy, precision, and recall levels. This study offers valuable contributions by shedding light on the appropriateness of various models for detecting and classifying faults in optical fiber.

Perspectives