Model Selection for Emission Models Based on Emission Factors

Transportation industry accounts for the majority of air pollutant emission, which is one of the leading factors of climate change. Many vehicle emission models have been proposed focusing on emission factors to study the pollutant emission issue. However, little research has been done in China on emission models, though China is a major air pollutant-emitting country. This paper first introduces two emission models- MOBILE model and MOVES model, then proposes model selection methods including AIC (Akaike information criterion) and BIC (Bayesian information criterion) to compare and select the better model which could be applied in China. Experimental results show that the value of AIC and BIC for MOVES model is significantly lower than MOBILE model, which implies that MOVES model has better performance on real data fitness and prediction. Our experimental findings may be useful for future research on air pollutant emission modelling in China.