Prediction of Baltic dry index with hybrid model: SARIMA-BSTS-Prophet

View at publisher Summarise with VIRAT
Authors: Wenhao Peng, Xinzhe Liu
Cited by: 0
References: 35

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

In recent years, the Baltic Dry Index (BDI) has witnessed significant volatility in the international dry bulk shipping market, effected by factors including shipping market supply, trade demand, and costs. In previous studies on BDI, different influencing factors and prediction methods were tested. However, researchers are persisting in finding more accurate methods to predict BDI. Therefore, the intention of the study is to identify the momentous factors resulting in BDI’s fluctuations and forecast its future trends. The hybrid forecasting method, proposed in this paper, combines the Seasonal Autoregressive Integrated Moving Average (SARIMA) model, Bayesian Structural Time Series (BSTS) model, and Prophet model. The proposed model based on Bayesian theory not only can achieve more accurate predictive results based on the characteristics of different influencing factors but also provide a more flexible model structure. The SARIMA-BSTS-Prophet model proposed can capture the residuals’ trend pattern of the international dry bulk shipping market. The relatively low MAPE value (3.54%) demonstrates that this method exhibits significant predictive performance, outperforming other long-term time series forecasting models and neural network forecasting models, and the prediction results are also more robust.

Fields of Study & Subjects

Forecasting Techniques and Applications Stock Market Forecasting Methods Time Series Analysis and Forecasting

Publication details

Journal / source
Maritime Policy & Management
Pages
1–32
Publication date
27 March 2026
Language
EN