Prediction of Pipeline Corrosion Rate Based on Cross-Validation Gradient Boosting Decision Tree

According to the idea of ensemble learning, a prediction model of pipeline corrosion rate was established based on the gradient boosting decision tree algorithm, and its hyper-parameters were optimized by grid search and cross-validation methods. The corrosion test data of a certain oil pipeline was utilized to verify the model and was compared with the results predicted by widely used BP neural network and support vector machine model. The results show that the average absolute percentage error of the gradient boosting decision tree model was 2.25%, which was lower than the 6.03% of the BP neural network and the 7.99% of the support vector machine. It indicated that the gradient boosting decision tree model had higher prediction accuracy and better generalization ability than the BP neural network and the support vector machine. In addition, this model had fine interpretability and could provide a new more practical method for the prediction of pipeline corrosion rate in the future.