Forecasting of meteorological elements time series and pricing of weather multi factor options: inte
Yongfeng Wei，Xiang Zheng
(Department of Statistics and Finance，School of Management，University of Science and Technology of China ,Hefei 230026，China)
With the increasing of abnormal climate change and frequent weather catastrophe happening in recent years, weather risk management has played a more and more important role in economic development. While weather derivatives are the most effective tools for management weather risk, how to price the weather derivatives is the core topic in this field. Current study aims to forecast meteorological elements time series and price of weather multi factor options by the integrating BP neural network and SARIMA model. We analyze the meteorological data from 1951to 2013 in Beijing base on the BP neural network and SARIMA model. There are two types of data in our analysis. One type is the monthly average temperature; the other type is the dynamic change of the average monthly rainfall. We use the data to calculate 4 weather indices: CDD (cooling degree days), HDD (heating degree days), EDD (energy degree days), CRI (cumulative rainfall index), and then we price some weather multi-factor options by using the estimates of meteorological elements time series of the underlying indices. Our results show that integrating the BP neural network and SARIMA model has strong nonlinear mapping ability, the forecast and valuation result is better than that of single model. Weather multi-factor options can analyze the different elements of weather in the economic effect which conceal in the weather option, meanwhile, it can avoid the weather risk effectively and gain profits.
Key words: weather multi factor option; weather option valuation; weather elements forecast; SARIMA model; BP neural network