The Study And Application Of The Simulation Of Huge Data Samples Based On The Approximation Theory

In all fields, such as the area of our common lifes, the scientific researches, and general economic situations, to discover the nature of the law, we have to face the large sample data which we analyze will help us perspect from the phenomenon. 0n the base of numerical approximation, this paper proposes the way of fitting a large sample data set, which is conducted by the least squares method with Pade approximation and Bezier curves.Whith the advantages of Pade approximation and Bezier curve in the numerical approximation, a general method of fitting a large sample data set has been raised in this paper. The paper is divided into four parts, the first part gives an proper introduction of large sample datas fitting methoda and theories, as well as the summarization and classification, from the polynomial interpolation of the earliest time to the radial basis function interpolation in recent time and related references. The second part describes the Pade approximation and illustrate how to fit a given set of large sample data with Pade approximation, and form which we are satisfied with the method. The third part begins with the Bezier curve, it also contends two parts:the first provides the process of using the least square method to fit a large sample data set and a detail analysis of the iterative calculation; the second discusses the Bezier curve in the fitting of the large population samlpe data set, and give us a subsequencial steps and clear detailed calculations. The fourth part give some examples of testing the Pade approximation method and Bezier curve method, at the time, compare the accuracy and point out necessary improvements.We get a general conculsion from the study of Pade approximation and Bezier curve in the fitting of a large sample data set and extend it to the more laege range of similar data sets fitting, as the large sample population data fitting. On both sides of the theory and pratice, it forms a effctive and completed method. Through the data test, we can see that the method can be used in various conditions to fit a large sample population data with a perfect accuracy and great potential for applications such as in demography, actuarial science and application of economic research.