METHOD FOR PREDICTING EMISSION AMOUNT OF SINGLE VEHICLE

摘要

A method for predicting emission amount of a single vehicle. A relevance vector machine model is established by using actual working condition features of a vehicle and corresponding emission data; a mutual information algorithm is introduced when the model is established, implementing screening of key features; and a frog leaping algorithm is introduced so as to optimize the calculation process of kernel function parameters of a relevance vector machine. By using the model, for a vehicle with an unknown emission level, only a small amount of random driving data needs to be acquired at low cost to serve as model input, and thus the model can directly output an emission result. The model is simple, convenient, and easy to implement.