Ship fuel consumption prediction model based on XGBoost algorithm

Ship energy consumption prediction is the basis and premise for ship energy efficiency assessment and optimization decision-making, which is of great significance for ship route speed design and ship energy efficiency optimization. Firstly, based on the ship's actual navigation records and environmental data, the characteristic factors which have a great influence on ship fuel consumption are extracted through correlation analysis, and these characteristic factors are taken as the input parameters of the model. Then, by filtering these characteristic factors based on data cleaning technology and relevant international standards, the sample data for modeling are obtained; and they are randomly divided into training and test samples according to the ratio of 0.8∶0.2. Then, the fuel consumption prediction model for the training samples is built based on XGBoost algorithm, and its accuracy is verified by the comparation of predictive value and true value of the test samples. Finally, the proposed model is compared with the neural network models, it is found that its decision coefficient can reach 0.967 and the running time is 2.723 seconds, the accuracy is almost the same as the neural network model but the running time is shortened by 70%. Therefore, the algorithm proposed in this paper is suitable for fast calculation of fuel consumption and real-time prediction of energy consumption in ship navigation.