A comparative study on prediction of throughput in coal ports among three models

作者：Liu Shuang; Tian Lixia; Huang Yuansheng^*

来源：International Journal of Machine Learning and Cybernetics, 2014, 5(1): 125-133.

DOI：10.1007/s13042-013-0201-5

• Least squares support vector machines
• Adaptive particle swarm optimization
• Coal port throughput

Three forecasting models, i.e., the least squares support vector machine (LSSVM), the neural network with back-propagation algorithm (BP), and a hybrid approach called APSO-LSSVM, are presented in this paper to predict the throughput of coal ports. A comparative study on the prediction accuracy among the three models is conducted. The purpose of this comparative study is to provide some useful guidelines for selecting a more accurate model to predict the throughput. The comparative results experimentally show that, in comparison with LSSVM and BP, the APSO-LSSVM has the more accurate accuracy and the better generalization performance regarding the indexes average error, mean absolute error and mean squared error.

• 出版日期2014-2
• 单位华北电力大学; 华北电力大学（保定）