Study on Weighting Function of Weighted Time Series Forecasting Model in the Safety System

The influence of observations always was ignored in traditional forecasting model when the parameter estimation was taken. A weighted time Series forecasting model has been established to emphasize the recent observations in safety system prediction. Based on safety system's characteristics that its weighting function should be a monotonous increasing function, the weighting function selection has be discussed in this article. Through analysis and forecasting, the weighting function which took time and residual into account at the same time is proved to have a higher accuracy than others. According to the characteristics of the system, the selection of the system of property right, weighted time series forecasting model can be applied to other time series system. It has practical significance.