Time Series Modeling And Forecasting Principles And Applications

Due to the time series data is not completely random, as well as easy accessibility, resulting in an increasing number of researchers it would be to find some of the laws and thus a specific time-series data for more accurate predictions. Time series analysis method is the study of statistics and economics, a hot spot, but also a recently developed quantitative forecasting methods, it is particularly applicable to economic class data analysis.This paper introduces the time-series nature of the various models and their corresponding features, and details of several basic models of time series. Box—Jeknins briefly describes the general theory of modeling Jeknins and methods. On this basis, the theory is applied to the actualTherefore, Based on time series analysis method to study the economics of time-series data, the principal tasks are as follows:1.For the Shanghai Composite Index data for ARMA modeling, using the optimal fixed-order method is relatively model-order determination and to estimate the parameters in the model, the final estimate the stock index (or yield) of the future value obtained ARMA model can be a good prediction of stock indexes, but the increase in step with the forecasting, forecast accuracy will drop, so we can only make short-term stock index forecasting.2. For containing the trend, seasonal time series data of economic classes, we use seasonal product model to model such data. Because this method removed the original data, trend and seasonal, so the model can better characterize the behavioral characteristics of such data, a more satisfactory prediction results so that decision-makers to develop sound policies to guide the sustainable development of China's economy.