Table 1

Overview of selected studies on container throughput forecasting models and their applicability to port operations

StudyMethodologyPorts/RegionKey contributionsLimitations
Farhan and Ong (2018) SARIMAVarious global portsCaptures seasonality in container dataLimited to linear patterns
Yang and Chang (2020) CNN-LSTM hybridEast Asian portsHigh accuracy in mixed precision settingsNo explainability tools
Kulshrestha et al. (2024) Decomposition + DL EnsembleMultinational datasetStrong predictive performanceHigh computational complexity
Xiao et al. (2023) Attention-based ensemble + XAIFour Asian portsIntegrates accuracy with partial explainabilityLimited model generalizability
Shen et al. (2025) Decomposed ensemble + XAIGate-in operationsHighlights role of interpretability in terminal logisticsFocuses on terminal not port-level
Rashed et al. (2018) Scenario-based hybrid modelingHamburg–Le Havre rangeIncorporates macroeconomic scenariosNo ML or XAI components
Xu et al. (2022) Comparative ML vs traditional modelsChinese portsBenchmarks ML vs ARIMA, emphasizes ML gainsNo feature interpretation provided
Source(s): Author’s own work

or Create an Account

Close Modal
Close Modal