LSTMモデルを用いたLNG船スポット運賃決定要因に関する研究

抄録

Liquefied Natural Gas (LNG) carrier spot freight rates are shaped by numerous external factors, including global economic conditions and seasonal fluctuations, resulting in heightened volatility. To address this variability, we identified the key drivers of spot freight rates for LNG carriers using market data and applied a Long Short-Term Memory (LSTM) model, widely recognized for its effectiveness in time-series forecasting. The model was trained on multiple variables, including LNG stock levels in importing countries, LNG market prices, newbuilding costs, vessel operating speeds, number of port calls, and spot freight rates. Our analysis revealed that the most influential predictors of freight rate variability were LNG inventories, LNG prices, vessel speeds, and number of port calls. These factors were determined to be critical in shaping LNG carrier spot freight rates. In contrast, the ‘newbuilding costs’ variable, often assumed to correlate strongly with freight rates, proved to be insufficient for forecasting LNG spot freight rates due to its weak predictive performance. The findings from this study are anticipated to improve the accuracy and dependability of shipping freight rate predictions, thereby providing valuable insights for decision-making in the maritime industry.