Time Series EMD Analysis And Prediction System Design And Implementation

Diagnosis and climate prediction is an important branch of atmospheric science, serve the national economic construction and national defense construction is an important project. By using time series analysis forecast, is still in the business climate is an effective method. But climate sequence is usually a random and non-stationary time series, brought difficult to forecast. In order to solve the problem of random sequence of tranquilization, people has been improved and explore the various processing scheme. This thesis describes a new method for time series processing_EMD method, theory and the key technology of decomposition, and based on the EMD eight mode of mean generating function sequence and the continuation of technology, the application of stepwise regression, the fitting error, rhythm and short, medium and long cycle of three different time scales of combination forecast method of modeling prediction.From analyzing the research background and significance of design in this thesis, we study the main content and structure; Then, are introduced in this thesis, design and development of the prediction system by use of four kinds of time series prediction method of mean generating function(stepwise regression, the fitting error of mean generating function related prediction method), rhythm, and the mathematical model and the implementation steps, considering the system USES the forecasting method and the system more products to provide to practical business applications, introduced the processing with the result of prediction for this two kinds of integration scheme and two scoring method. Then, from the perspective of current situation, background and demand, designed a certain characteristic of time series analysis and forecasting. Then, this thesis introduces to build in this thesis, the design of the system should have what kind of function module and the operation process and business platform operation process; Realize the EMD decomposition was introduced in detail the key problems and technical difficulties, and in the EMD decomposition is given in the specific process and the core code, and then compare the result calculated by example and several EMD to discuss the reliability of the system the EMD decomposition., then puts forward the EMD with visual c # computer language theory to build the more powerful analysis and prediction of time series business platform and to realize the general requirements of the prediction system and implement of the specific steps. Finally, an example for operating, combining with the system of single time series prediction and trend prediction, and the implementation process of the two core functions such as all the features of steps of the system are analyzed in a comprehensive test.In this thesis, design and development of system, is a powerful and easy to operation of time series analysis and prediction system. Have thesiss into practical application, the system background analysis and prediction to climate provides a new technical support, provide meteorological services for more reference. But application indicates that in the climate prediction, and many other climate prediction method, there are still certain instability. This gives research has brought more challenges, more research space. To conquer the nature may never end.