The model used for gasoline engine control system should have the required accuracy and be as simple as possible. Generally, the model is described by a polynomial, but the number of regressors tends to be enormous in a gasoline engine with a non-linearity. Conventional engine modeling has not established a systematic method for reducing the regressors. Therefore, in this study, the effect was verified using the Lasso regression, which is known as an effective regularization method for pruning

more »

... egressors. As a result, the model accuracy was improved by considering the physical characteristics. It is also found that there is a possibility that unknown important physical properties can be extracted from the measured data.