Overview of selected studies on container throughput forecasting models and their applicability to port operations
| Study | Methodology | Ports/Region | Key contributions | Limitations |
|---|---|---|---|---|
| Farhan and Ong (2018) | SARIMA | Various global ports | Captures seasonality in container data | Limited to linear patterns |
| Yang and Chang (2020) | CNN-LSTM hybrid | East Asian ports | High accuracy in mixed precision settings | No explainability tools |
| Kulshrestha et al. (2024) | Decomposition + DL Ensemble | Multinational dataset | Strong predictive performance | High computational complexity |
| Xiao et al. (2023) | Attention-based ensemble + XAI | Four Asian ports | Integrates accuracy with partial explainability | Limited model generalizability |
| Shen et al. (2025) | Decomposed ensemble + XAI | Gate-in operations | Highlights role of interpretability in terminal logistics | Focuses on terminal not port-level |
| Rashed et al. (2018) | Scenario-based hybrid modeling | Hamburg–Le Havre range | Incorporates macroeconomic scenarios | No ML or XAI components |
| Xu et al. (2022) | Comparative ML vs traditional models | Chinese ports | Benchmarks ML vs ARIMA, emphasizes ML gains | No feature interpretation provided |
| Study | Methodology | Ports/Region | Key contributions | Limitations |
|---|---|---|---|---|
| SARIMA | Various global ports | Captures seasonality in container data | Limited to linear patterns | |
| CNN-LSTM hybrid | East Asian ports | High accuracy in mixed precision settings | No explainability tools | |
| Decomposition + DL Ensemble | Multinational dataset | Strong predictive performance | High computational complexity | |
| Attention-based ensemble + XAI | Four Asian ports | Integrates accuracy with partial explainability | Limited model generalizability | |
| Decomposed ensemble + XAI | Gate-in operations | Highlights role of interpretability in terminal logistics | Focuses on terminal not port-level | |
| Scenario-based hybrid modeling | Hamburg–Le Havre range | Incorporates macroeconomic scenarios | No ML or XAI components | |
| Comparative ML vs traditional models | Chinese ports | Benchmarks ML vs ARIMA, emphasizes ML gains | No feature interpretation provided |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.