Skip to Main Content
The prediction problem of first typhoon has been studied in the meteorological field. The reason is that accurate prediction helps in early stage prevention and decreasing economic loss. But the appeared dates of first typhoon are uncertain every year. Grey system theory is one of the methods that are used to study uncertainty, it is superior in mathematical analysis of systems with uncertain information. The change of first typhoon can be view as a grey process, therefore grey system theory can be used to solve the prediction problem. Since the change of first typhoon is affected by variant random factors, it is difficult to obtain high predicted accuracy by single grey model (GM). Therefore, we presented statistical methods to improve the predicted accuracy of GM. In this paper, we proposed a hybrid grey-based model to predict the first typhoon over the South China Sea. The work procedure: First, GM(1,1) is applied to predict the appeared dates of first typhoon. Second, second order polynomial regression is integrated into GM(1,1) to improve predicted accuracy. This new generated model is defined as RGM(1,1). Finally, Markov-chain is integrated into RGM(1,1) to enhance the predicted accuracy further. This new generated model is defined as MRGM(1,1). The experimental results show that the proposed hybrid MRGM(1,1) model has proved an effective tool in first typhoon prediction problem.