|
In the reality of economic life, due to climatic conditions, production conditions, holidays and the living, customs and other factors, making the real economy often repeated annual regular cycle changes, and annual changes in the law are broadly similar this change is the change of season we often say, sometimes also known as cyclical changes. In fact, the vast majority of economic time series data contain seasonal changes in composition, many researchers to study economic phenomena with seasonal changes, but most researchers seasonal variation component in the economic data as \, the focus of how to removed the seasonally adjusted time series; or focused on the seasonal time series prediction. However, in many cases, we need to know the size of the seasonal changes in the economic phenomenon, so in different seasons will be able to take the appropriate measures to deal with these seasonal fluctuations, seasonal index is a measure of seasonal changes in the size of the index. Study how to more accurately estimate the index of the seasons, both in theory have a higher value, or in practical applications. The basic idea is the seasonal index quarterly average method, the the seasonal index estimation method is studied in this paper is based on the idea of ??the quarterly average of Law commenced. However, economic data obtained in real economic life, the variety, the quarterly average in many conditions that can not be directly applied; the same time, the seasonal index quarterly average method is just a description of the statistical , and describe the value of statistics in the analysis of economic statistics is very limited, so you must study the seasonal index inferential statistics. In this paper, on the basis of the quarterly average of Law, proposed three estimation methods for different data types seasonal index for each estimation method are given the seasonal index of inferential statistics. The first method and the second method is the use of seasonal dummy variable model to estimate the seasonal index, seasonal dummy variable model parameters estimated indirect estimates of the seasonal index, these two methods can not only get the point estimate of the seasonal index can get the seasonal index interval estimation, seasonal changes and also the existence of economic time series data inference test. In both methods, the former suitable for continuous years of economic data, while the latter is suitable for shorter observation period panel data, these two methods are the use of the idea of ??regression to estimate and analyze the seasonal index. The third method uses sampling estimation method to estimate the seasonal index, the direct use sample data to estimate the overall seasonal index, greatly reduce the data requirements, thus a wider range of applications.
|