曲线拟合
已知离散点上的数据集
,即已知在点集
上的函数值
,构造一个解析函数(其图形为一曲线)使
在原离散点
上尽可能接近给定的
值,这一过程称为曲线拟合。最常用的曲线拟合方法是最小二乘法,该方法是寻找函数
使得
最小。
MATLAB函数:p=polyfit(x,y,n)
[p,s]=
polyfit(x,y,n)
说明:x,y为数据点,n为多项式阶数,返回p为幂次从高到低的多项式系数向量p。x必须是单调的。矩阵s用于生成预测值的误差估计。(见下一函数polyval)
多项式曲线求值函数:polyval( )
调用格式: y=polyval(p,x)
[y,DELTA]=polyval(p,x,s)
说明:y=polyval(p,x)为返回对应自变量x在给定系数P的多项式的值。
[y,DELTA]=polyval(p,x,s) 使用polyfit函数的选项输出s得出误差估计Y
DELTA。它假设polyfit函数数据输入的误差是独立正态的,并且方差为常数。则Y DELTA将至少包含50%的预测值。
=========================
练习:如下给定数据的拟合曲线,x=[0.5,1.0,1.5,2.0,2.5,3.0],
y=[1.75,2.45,3.81,4.80,7.00,8.60]。
解:MATLAB程序如下:
x=[0.5,1.0,1.5,2.0,2.5,3.0];
y=[1.75,2.45,3.81,4.80,7.00,8.60];
p=polyfit(x,y,2)
x1=0.5:0.05:3.0;
y1=polyval(p,x1);
plot(x,y,'*r',x1,y1,'-b')
计算结果为:
p =0.5614 0.8287
1.1560
即所得多项式为y=0.5614x^2+0.8287x+1.15560
另:matlab
polyval与polyvalm的区别是什么?
在坛子里发个网址真费劲,感谢原帖的几位大侠。
http://www.ilovematlab.cn/viewthread.php?tid=2538
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
polyval 只是计算一个多项式,这个你应该明白吧,help里的例子很清楚:
就是把x值带进去。
polyvalm比较特别的。但是明白这个之前,你需要明白什么是 characteristic
polynomial:
In linear algebra, one associates
a polynomial to
every square matrix,
its characteristic
polynomial. This polynomial encodes several important
properties of the matrix, most notably
its eigenvalues,
its determinant and
its trace.
Suppose we want to compute the characteristic polynomial of the
matrix
We
have to compute the determinant of
and
this determinant is
The
latter is the characteristic polynomial
of A.
现在回到我们自己的问题, 对于Y = polyvalm(p,X), 他的计算方法是:
Y = P(1)*X^N + P(2)*X^(N-1) + ... + P(N)*X + P(N+1)*I
如果写出for循环是:
for i = 1:np %多项式长度,m是X的维数
Y = X * Y +
diag(p(i) * ones(m,1));
end
这个循环比较简单,不要我解释吧?
举个例子给你看,我把循环拆开:
复制内容到剪贴板代码:
p=[1 2 3]
X=[1 2; 3 4]
np = length(p);
[m,n] = size(X);
Y = zeros(m,m)
i = 1
Y = X * Y +
diag(p(i) * ones(m,1))
i = 2
Y = X * Y +
diag(p(i) * ones(m,1))
i = 3
Y = X * Y +
diag(p(i) * ones(m,1))
你再运行:
p=[1 2 3]
X=[1 2; 3 4]
polyvalm(p,X)
(By
math)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
总结一下:
polyval(p, t)是计算出来:
p(1)*t^n + p(2)*t^(n-1) + .... + p(n)
polyvalm(p, X)是计算出来:
p(1)*X^n + p(2)*X^(n-1) + .... + p(n)*I
这里要注意最后的I就是单位阵,不要忘了,否则会出错,比如:
>> p=[1 2 3];
>> X=[1 2; 3 4];
>> polyvalm(p, X)
ans =
12 14
21 33
>> p(1)*X*X+p(2)*X+p(3)
ans =
12 17
24 33
>> p(1)*X*X+p(2)*X+p(3)*eye(2,
2)
ans =
12 14
21 33
===============================
P =
POLYFIT(X,Y,N) finds the coefficients of a polynomial P(X) of
degree N
that fits the data Y best in a least-squares sense. P is a
row vector
of length N+1 containing the polynomial coefficients in
descending
powers, P(1)*X^N + P(2)*X^(N-1) +...+ P(N)*X + P(N+1).
[P,S] =
POLYFIT(X,Y,N) returns the polynomial coefficients P and a
structure S
for use with POLYVAL to obtain error estimates for
predictions. S contains fields for the triangular
factor (R) from a QR
decomposition of the Vandermonde matrix of X, the degrees of
freedom
(df), and the norm of the residuals
(normr). If the data Y are random,
an estimate
of the covariance matrix of P is (Rinv*Rinv')*normr^2/df,
where Rinv
is the inverse of R.
[P,S,MU] =
POLYFIT(X,Y,N) finds the coefficients of a polynomial in
XHAT =
(X-MU(1))/MU(2) where MU(1) = MEAN(X) and MU(2) = STD(X).
This
centering
and scaling transformation improves the numerical properties
of both the
polynomial and the fitting algorithm.
Warning
messages result if N is >= length(X), if X has
repeated, or
nearly
repeated, points, or if X might need centering and scaling.
Class
support for inputs X,Y:
float: double, single
给个例子。
clc;clear
x=-1:0.1:1
for k=1:length(x)
y(k)=x(k)^2+0.1*rand
end
[p,s]=polyfit(x,y,2)
s.R
plot(x,y,'o',x,polyval(p,x))
结果:
p =
1.0072 0.0091 0.0439
s =
R: [3x3 double]
df: 18
normr: 0.1153
ans =
-2.2509 -0.0000 -3.4208
0 2.7749 0.0000
0 0 -3.0492
s是个结构数组。
S contains fields for the triangular factor (R) from a QR decomposition of the Vandermonde matrix 【范德蒙特矩阵】of X, the degrees of freedom (df)【自由度】, and the norm of the residuals (normr)【范数】.