Software design of sand-dust strom warning system based on grey correlation analysis and particle swarm optimization support vector machine

This paper proposes the design concept of sand-dust storm warning system based on data collection techniques, grey relation analysis and support vector machine classification forecasting model. We have successfully designed sand-dust storm warning system by the three sub-modules that they are the analysis, classification and forecasting system, and the article details grey relation analysis, support vector machine classification and forecasting module. According to the characteristics of sand-dust storm and the advantages of support vector machine (SVM) in solving learning problems with fewer samples, sand-dust storm classification and forecasting model based on PSO-SVM is proposed. We use the grey relation analysis to find the several major factors of sandstorm happening. The input of model is the major factors of sandstorm happening, and the output results are sand-dust storm grade and forecasting results for aim to achieve sand-dust storm early warning.