Table_1_Analysis of influencing factors and prediction of China’s Containerized Freight Index.xlsx

China, as a major maritime nation, the China Containerized Freight Index (CCFI) serves as an objective reflection of the Chinese shipping market and an important indicator for understanding China’s shipping industry globally. The shipping market is a complex ecosystem influenced by various factors, including vessel supply and demand, cargo supply and demand relationships and prices, fuel prices, and competition from substitute and complementary markets. To analyze and study the state of the Chinese shipping market, we selected the CCFI as an indicator and collected data on six factors that may affect the overall shipping market. These factors include “ the China Coastal Bulk Freight Index(CCBFI)”, “the Baltic Dry Index(BDI)”, “the Yangtze River Container Freight Index”, “Global: Aluminum (minimum purity of 99.5%, London Metal Exchange (LME) spot price): UK landed price”, “Major Ports: Container Throughput”, and “Coal Price: US Central Appalachia: Coal Spot Price Index”. Then, we constructed an analyticaland predictive framework using Deep Neural Network (DNN), CatBoost regression model, and robust regression model to study the CCFI. Based on the R2 results of the three models, it is evident that DNN provides the best analytical and predictive performance for the CCFI, accurately forecasting its changes. Additionally, the robust regression model indicates that “Global: Aluminum (minimum purity of 99.5%, LME spot price): UK landed price” has the greatest impact on the CCFI. Finally, from a business perspective, we provide some suggestions for China’s container shipping industry.

作为海洋大国，中国出口集装箱运价指数（China Containerized Freight Index, CCFI）客观反映了中国航运市场的运行状况，亦是全球范围内研判中国航运产业发展态势的重要指标。航运市场是一个复杂的生态系统，受多重因素影响，涵盖船舶供需状况、货物供需关系与价格、燃油价格，以及替代与互补市场的竞争格局。为分析研判中国航运市场的运行状态，本研究选取CCFI作为核心指标，并收集了六项可能影响航运市场整体走势的影响因子数据，分别为：中国沿海散货运价指数（China Coastal Bulk Freight Index, CCBFI）、波罗的海干散货指数（Baltic Dry Index, BDI）、长江集装箱运价指数、全球铝（最低纯度99.5%，伦敦金属交易所（London Metal Exchange, LME）现货价格）：英国到岸价、主要港口集装箱吞吐量、以及美国阿巴拉契亚中部煤炭现货价格指数。随后，本研究采用深度神经网络（Deep Neural Network, DNN）、CatBoost回归模型与稳健回归模型构建了分析与预测框架，用于CCFI的相关研究。基于三款模型的决定系数（R²）结果可知，深度神经网络在CCFI的分析与预测任务中表现最优，可精准预判其波动变化。此外，稳健回归模型的结果显示，全球纯度不低于99.5%的铝（伦敦金属交易所现货价格）：英国到岸价对CCFI的影响程度最大。最后，本研究从产业实务视角出发，为中国集装箱航运业提出了若干发展建议。