基于多算法与XGBoost融合的柴油机瞬态排放预测OA

To address the limitation of single algorithm,a novel hybrid model combining multiple machine learning algorithms,including artificial neural network(ANN),support vector machine(SVM),nonlinear autoregressive with exogenous inputs(NARX),long short-term memory networks(LSTM),gated recurrent unit(GRU),transformer and temporal convolutional networks(TCN),with extreme gradient boosting(XGBoost)was proposed to predict transient emission from diesel engines.Additionally,random forest algorithm was used to select best input variables of the hybrid model,and particle swarm optimization(PSO)and genetic algorithm(GA)algorithms were employed to determine the optimal hyperparameters.The results indicate that the hybrid model integrates all advantages of different machine learning algorithms and achieves more precise prediction.Meanwhile,the hybrid model demonstrates excellent prediction performance on both the training dataset and validation dataset,with R2 values exceeding 0.980 and 0.967,respectively.In the test dataset,the R2 value of the hybrid model is above 0.930,which demonstrates good generalization.