第三章数据拟合B.ppt

 3 Data Approximating Mehtod 3 数据拟合方法 Useful Methods： Conclusion Main content of data approximating: From the gven huge data, get a ‘model’, that is, to find a approximating function to very close to the data such that the local jump small enough. How to construct the fit function? Least-Squares (L-S) method 最小二乘法 The method was first described by Carl Friedrich Gauss around 1794 at the age of eighteen. Legendre was the first to publish the method in 1805, however. Gauss did not publish the method until 1809. Least squares can be interpreted as a method of fitting data. The best fit, between modeled and observed data, in the least-squares sense is that instance of the model for which the sum of squared residuals has its least value. (即以残差平方和最小问题的解来确定拟合函数的方法.) A residual is the difference between an observed value and the value given by the model. The method of least squares is used to approximately solve overdetermined systems, i.e. systems of equations in which there are more equations than unknowns. Least squares is often applied in statistical contexts, particularly regression analysis. Due to the observed data，we can choose power function(幂函数)、exponent function(指数函数)、sinx or cosx and so on. Sum of the squared residual vector(残差向量的各分量平方和): If mn+1，then the system is called by overdetermined equations(超定方程组（方程个数未知数个数）). Remark：具体用几次多项式拟合，可据实际情况而定。可先画草图，将已知点描上去，看与什么函数相近，就以什么函数拟合. Example 2 考虑下表给出的离散点 我们看到，给定的离散数据在一条直线 附近，因此，用直线来拟合这些数据是合 适的，据此，我们确定拟合模型为： 这时，基函数为： ， ，用 于拟合的函数类为： ，取权： 经计算得到： 利用这些值，得到正规方程为： 此方程组的解为： ， 故解出的连续模型为： 前面的表中最后一列给出了 的值，可 与 的值相比较。 结果见下图： Example 3 考虑下表给出的离散点： 这次，我们用二次曲线来拟合，我们有三 个基函数： ， 拟合模型 为： 取： 依据 可以算出： 正规方程为： 求解得到： 因此所求的模型为： 最小平方残差（各点拟合残差的平方和） 求解线性最小二乘拟合的步骤 (1) 根据已知的离散点 ，画出这些 点，观察其分布，选择合适的拟合模型， 即选择用直线，二次曲线，三次曲线等去 拟合； (2) 计算 ，从而求出 ； (3) 计算