Time Series Analysis System

In many domains, such as the natural sciences, social sciences, engineering technology and so on, the random phenomenon, which is characteristic of probability and happens in order by timer, is widespread. This phenomenon becomes analyzable randomized data, if it is recorded by observing. We usually call it time series, means a kind of orderly random data. The time series analysis provides a method-set of dynamic data processing; this method is mainly used to all kinds of data with the corresponding mathematical model to approximate description. Through the model analysis, can be more essential understanding the data's inherent structure and complex characteristic, according to forecast the developing trends and achieve the necessary control.At present, the diversity of time series analysis algorithm, make it difficult to choose the appropriate algorithm, more easily choose the wrong or inappropriate model to result in the inaccurate or even invalid forecast. What kinds of model shall we choose usually based on checking or expertise knowledge. Although there are many time series analysis software, but more complex in operation, many users restricted by this, furthermore, due to the software application limit in the time series analysis, lead to the wrong results, result in the wrong judgment. Therefore, develop an application which is user-friendly, fulfils all the users requirements is very significance.Firstly, in this paper, according to the characteristics of time series analysis and analysis system, Organized the composition and suitable function in the time series analysis system. Secondly The high-level design of the system, and frame conjunction with the application of the basic object-oriented system model and the model / view / control procedures (MVC, Model / View / Controller), proposed system design principles and objectives, and gives solutions by modular. System design is divided into three Layers UI Layer, Facade Layer and Engine Layer. All levels are relatively independent of the framework, coupled with a lower degree. Then use UML modeling language to describe the design.In this paper, based on my research of time series analysis algorithm, achieve several time series analysis methods, and test the effectiveness of the algorithm through data validation. Finally design time series analysis system interface, the main function modules details and the realization of the current system. In this paper the time series analysis system combines expert system, virtual instrument technology, with the participation of experts in the analysis of time series, also display the dynamic changes process, the systematic analysis of the time series, we can see the picture and the details.