The Application Research Of The Optimal Estimation And Wavelet Analysis Theory In The Economic Analysis

While economy globalizing and finance incorporating, the Chinese economy will join in the world economy system faster and wider. But at present, the Chinese economy is in the special period of the system changing. Then many practical problems need the corresponding new theory. Therefore it is very necessary to explore the methods and techniques, which are more systematic, realistic, real-time and dynamic, and more adaptable to the Chinese economy.In this paper, I use the Kalman filtering, the strong tracking filtering and the multiresolution analysis methods to analyze some economic phenomena. Further, I compare the new methods with the traditional methods by some examples. Then I gain some meaningful research results. The main contributions of this paper are as follows:1. The B is an important index which can weigh the asset's systemic risk. But the B in the stock market of China is stochastic and dynamic to a certain extent, due to the disturbance of the various uncertain factors. In this paper, I obtain the dynamic estimation of the B based on the Kalman filtering method. Then it is reported that the results of this method is more effective compared with the traditional regression method. It clearly demonstrates the advantage of the Kalman filtering method.2. Some domestic scholars pay attention to the information content of dividends in the stock market of China. In this paper, I study this problem by the correlation between the dividends and the permanent earnings. Firstly, the strong tracking filtering method allows me to obtain the dynamic estimation of the permanent earnings. It is found that this method has the strong ability to track the permanent earnings. Then I analyze the correlation between the dividends and the permanent earnings. The results show that there is no significant correlation between them. So it is implied that dividends can't reflect the future earning information nowadays.3. Since there are a lot of periodic non-stationary stochastic processes, I combine the Kalman filtering and multiresolution analysis methods, and propose thewavelet-Kalman filtering hybrid estimating and forecasting method. This method has the real-time, recursion and multiresolution characteristics. In this paper, using this method, I realize the real-time tracking and dynamic multistep forecasting in one cycle. Then this method is tested by the real economic data. It is clearly that this new method is adaptable to the economy real-time forecasting and multiresolution analysis.