基于GBDT的船舶油耗预测模型设计

Prediction model of ship fuel consumption based on GBDT

建立准确的船舶油耗预测模型是实现船舶节能减排的基础。文中以某一远洋船舶作为研究对象,对船舶的实际运营数据进行分析和预处理,结合GBDT算法构建船舶油耗的预测模型。将模型在测试数据集上的预测值与实测数据进行对比,结果表明预测误差低于6.98%,优于随机森林和决策树的预测结果。该文验证了基于GBDT所建立的油耗预测模型的准确性和实用性,对于后续船舶航行优化措施具有重要意义。

Establishing an accurate prediction model of ship fuel consumption is the basis for realizing ship energy saving and emission reduction.This paper took an ocean-going vessel as the research object,preprocessed the actual operating data of the ship,and built a prediction model of ship fuel consumption combined with the GBDT algorithm.Comparing the predicted value of the model on the test data set with the actual measured data,the results showed that the prediction error was less than 6.98%,which was better than the prediction results of random forest and decision tree.This paper verified the accuracy and practicability of the fuel consumption prediction model based on GBDT,which was of great significance for the subsequent measures to optimize ship navigation.