第4讲 曲线拟合.pptVIP

* * * * * * * 引入记号 定义加权内积 * 矩阵形式(法方程组)为 方程组式化为 * 平方误差为 作为特殊情形,用多项式作拟合函数的法方程组为 * Subject: What Weighted-Least-Squares Fitting capabilities are available in MATLAB 6.1 (R12.1) and the Toolboxes? Problem Description: Currently, the presence of data outliers can create an undesirable fit. Because the outlier lies far away from the true pattern of data, it induces error to the true fit. A workaround to this problem would be to minimize the weight(s) of such outlier(s). Solution: In MATLAB, the LSCOV function can perform weighted-least-square regression. x = lscov(A,b,w)where w is a vector length m of real positive weights, returns the weighted least squares solution to the linear system A*x = b, that is, x minimizes (b - A*x)*diag(w)*(b - A*x). w typically contains either counts or inverse variances. In addition, there are three toolboxes you can use to implement weights for your fits: ================== 1. Statistics Toolbox: ================== Weighted linear regression in the Statistics Toolbox is part of the ROBUSTFIT function, B = ROBUSTFIT(X,Y,WFUN,TUNE,CONST) uses the weighting function WFUN and tuning constant TUNE. WFUN can be any of andrews bisquare, cauchy, fair, huber,logistic, talwar, welsch. As an alternative to specifying one of the named weight functions shown above, you can also write your own weight function (wfun) that takes a vector of scaled residuals as input and produces a vector of weights as output. For documentation on ROBUSTFIT, you can type doc robustfit (without quotes) at the MATLAB command prompt or view the online documentation found at the URL below: /help/toolbox/stats/robustfit.html For MATLAB versions prior to 7.1 (R14SP3), we do not support a non-linear weighted least-square fit in the Statistics Toolbox. In MATLAB 7.1 (R14SP3), the demo Weighted Nonlinear Regression, addresses this and is also available on the web at the following link /products/statistics/demos.html?file=/products/demos/shipping/stats/wnlsdemo.ht