Comparative Study of Traffic Carbon Emission Prediction Models Based on X-KELM

Transportation is an important industry for carbon emission reduction, and swarm intelligence technology, as a very important and significant branch of artificial intelligence, is an efficient method for solving complex optimization problems, and shows better performance in the field of carbon emission prediction. In this paper, we use the 2015–2019 Shanghai transportation carbon emission data to do example analysis, and apply Bundle Adjustment (BA), Firefly Algorithm (FA), Cuckoo Search (CS), Grey Wolf Optimizer (GWO) and Whale Optimization Algorithm (WOA) to KELM to establish a transportation prediction carbon emission model. From the data conclusions, it can be seen that the BA-KELM model is more stable, with the lowest coefficient of variation compared to the lowest, with a value of 0.065, and the minimum number of iterations of WOA-KELM is 6 times and the highest model accuracy, with the values of MAE, RMS, MAPE, and R2 of 0.0257, 0.0331, 0.1070, and 0.9686, respectively. it can be seen that from the data conclusions, the WOA -KELM model has high accuracy and fewer iterations in transportation carbon emission prediction, but the stability needs to be improved.