Discovery model and classification prediction of dust pollution at different altitudes based on LS- SVM and BP algorithm

Abstract

In order to solve the classification prediction of dust pollution at different altitudes, the least square support vector machine(LS-SVM) and BP neural network is used to construct the distribution model. Built by LS-SVM, the accuracy of the model was verified by BP neural network with the realtime dust pollution data of different high monitored by Unmanned aerial vehicles. The data analysis shows that the dust pollution below 30 meters is much serious than that above 90 meters, which mainly concentrated in the low altitude area of 20 to 30 meters. While in the range of 80 to 250 meters, the dust pollution is basically proportional to the height, that is to say, the higher the altitude, the greater the pollution index.