An ICAKELM-Based Model to Predict Container Throughput of Ports

A model integrating multiple processing technologies is developed to improve the accuracy and the consistency of the prediction of the container throughput of a port.The ICEEMDAN is used to resolve a port's container throughput sequence and analyze the complexity of the sub-sequences.The complexity of subsequences is examined with sample empathy.The ARIMA(autoregressive integrated moving average) and the ICAKELM(KELM optimized with imperialist competitive algorithm) are used to do subsequence prediction respectively.The ICAKELM is used to do non-linear composition of all the subsequences and find the final prediction.Experiments show that the proposed method gives better short-term prediction than traditionally sole BP or ARIMA.