Prediction of Optical Fiber Cable Lifespan Based on Bi-LSTM and Attention Mechanism

We proposed a novel method for predicting the service life of optical cables based on Bi-LSTM combined with the Attention Mechanism. This method uses Bi-LSTM (Bidirectional Long Short-Term Memory) to extract features of optical cable length and introduces the Attention Mechanism to calculate the weight of each time node, making the model more accurate in predicting the service life of optical cables. Meanwhile, a prominent feature of the system is utilizing various inputs derived from optical cables parameters under all-weather condition. Simulation results show that the system can accurately predict the service life of optical cables, providing a more reliable and efficient method of optical cable management and maintenance for related industries.