A novel approach based on the Gauss-vSVR with a new hybrid evolutionary algorithm and input vector decision method for port throughput forecasting

摘要: The prediction of port throughput is very complicate, and its accuracy affected by many socio-economic factors, particularly their embedded distributed randomness these factors mixed noises produced in the processes data collection, transformation, calculation. Firstly, view v-support vector regression hybridized with Gauss function (briefed as Gauss-vSVR model), to well solve nonlinear noises, this paper uses model simulate evolving system series. Then, look for more suitable parameter combination take into account that GA still suffers from problems trapped local optima time-consuming, study integrates global chaotic perturbation algorithm using Cat mapping acceleration search employing cloud theory, i.e., abbreviated genetic (CCGA), determine values an model. Additionally, based on principal component analysis correlation method, input decision method (namely IVD) proposed identify final variables Finally, hybridization IVD CCGA model, namely IGvSVR-CCGA, forecasting. Subsequently, associate two largest Chinese ports, Shanghai Port Tianjin Port, are employed practical examples test forecast performance. numerical results indicate hybrid forecasting receives satisfied performance than other classical models; meanwhile, also obtains higher optimal efficiency alternative algorithms.