港口吞吐量的马氏链-时序分析预测

摘要

A novel approach of port throughput scientific forecast was studied by a comparative analyzing between the time series analysis (TSA) and Markov chain method. TSA and the Markov chain were both used to make prediction according to nearly twenty years historical throughput data of Wenzhou port. A composite model, which was made up by TSA and Markov chain, was also applied to make prediction of port throughput, and put Markov chain into a role of calibration mechanism. Results show that prediction accuracy of composite model is significantly improved by 50% than TSA, dramatically improved by 75% than single Markov chain. Markov chain-TSA forecasting model was established based on the results of verifying actual throughput data. The model can both reflect the growth trend and random properties of fluctuations in the time series of port throughput. The model is more in line with the actual throughput of the port changes.%通过比较分析时间序列分析(TSA)和马尔科夫链预测方法,研究港口吞吐量科学预测的新方法.组合温州港近20多年的历史吞吐量数据,分别采用TSA和马尔科夫链进行预测.将TSA与马尔科夫链校正模型相结合,进行港口吞吐量预测.结果表明,上述复合模型较之TSA模型平均预测精度提高50％,较之单一的马尔科夫链平均预测精度提高75％.根据吞吐量实际数据的验证结果,建立马氏链-时序分析预测模型.结果表明,该模型能够同时反映吞吐量序列的增长趋势和随机波动性,更符合港口吞吐量的实际变化情况.