A load prediction method for data transmission in optical communication networks based on LSTM Model

The massive transmission of data in optical communication networks leads to an increase in the load of optical communication channels, making it more difficult to predict data transmission loads. A data transmission load prediction method based on the LSTM model is proposed for optical communication networks. We constructed a data transmission channel model for optical network communication, calculated the data transmission rate, and combined the transmission efficiency calculation results with the joint feature selection method to obtain the data transmission load sequence. Then an LSTM prediction model was built and it was optimized using a particle swarm optimization algorithm. The obtained data transmission load sequence into the LSTM model was input to achieve data transmission load prediction. The experimental results show that the average accuracy of data transmission load prediction in the optical communication network using this method is 98.9%, the average prediction time is 0.64 seconds, and the highest service rejection rate is only 14.5%.