End-to-end price prediction model for CAMIO

Abstract

In this study, we present a machine learning–driven framework for end-to-end price prediction within the CAMIO logistics platform. We employ XGBoost regression to model the inherently non-linear relationships among trip distance, truck type, and geographic zones. Through gradient boosting of decision-tree ensembles—with integrated regularization and sparsity awareness—our approach addresses the limitations of the prior rule-based algorithm. On a held-out test set, the XGBoost model achieved a root-mean-square error (RMSE) of 2.995.02, a mean absolute error (MAE) of 1.776.69, and an R² of 0.94, reflecting substantial gains in estimation accuracy and resilience to outliers. Moreover, the model’s feature-importance metrics confirm that distance and regional factors are the primary drivers of price variation, offering actionable insights for operational planning. These results underscore the efficacy of gradient-boosted tree models in complex pricing environments and suggest a clear path toward enhanced customer satisfaction and operational reliability for CAMIO.