Short Term Forecast of Container Throughput: New Variables Application for the Port of Douala

An accurate container throughput forecast is vital for any port. Since overall improvements in port performance and competitiveness can be derailed by bottlenecks, ports need to find leverage identify prioritize measures improve weak key indicators (KPIs) attain growth opportunities. Prior studies had modeled from socio-economic projection factors. This study aims provide a practical method forecasting the optimal physically handle/attract given certain level of terminal operation efficiency through random forest (RF) multilayer perceptron (MLP) models. The variables are derived operations dimension include ship turnaround time, vessel draft, dwell berth productivity, storage capacity, custom declaration time. Evaluations made based on R-squared, NRMSE, MAE MAPE. Model comparison deduced with seven competing models forecasting. findings indicate that RF model potential candidate engineering Douala interpretation provided feature importance partial dependence plots. this will help reduce uncertainty management spot bottlenecks engage better planning development projects which strengthen their international competitive advantage.