论文部分内容阅读
The ultimate solution of the problem "Which smoothing method is better" is offered.A method of noise filtering based on confidence interval evaluation is described.In the case of the approximation of a function by least-squares method, the minimal confidence interval is used as a criterion for the selection of the proper parameters of the approximating function.In the case of the polynomial approximation optimized parameters include the degree of the polynomial, the number of points (window) used for the approximation, and the position of the window center with respect to the approximated point.The special considerations on confidence interval evaluation and quality of polynomial fit using noise properties of the data array are discussed.The Method provides the lowest possible confidence interval for every data point.