E********r
发帖数: 99 | 1 请问：如何用least square approach解超定线性方程组？
Use a least squares approach to solve overdetermined system of
linear equations?
用MATLAB解很容易，一下就出来了，但是怎么用C的代码解呢？
先谢了！ |
b*****y
发帖数: 163 | 2 You can use subroutines in CLAPACK.
Dense LS-problem is usually solved using:
1. Normal equation
2. QR factorization
3. SVD
【在 E********r 的大作中提到】
: 请问：如何用least square approach解超定线性方程组？
: Use a least squares approach to solve overdetermined system of
: linear equations?
: 用MATLAB解很容易，一下就出来了，但是怎么用C的代码解呢？
: 先谢了！
|
E********r
发帖数: 99 | 3 Thank you.
CLAPACK seems to be designed in Linux? I don't
know how to make it work in Windows.
I found a book 'numeical recipes in C'. Maybe
I will use SVD now.
【在 b*****y 的大作中提到】
: You can use subroutines in CLAPACK.
: Dense LS-problem is usually solved using:
: 1. Normal equation
: 2. QR factorization
: 3. SVD
|
b*****y
发帖数: 163 | 4 There is a win32 version of clapack.
You can check them out from netlib.
【在 E********r 的大作中提到】
: Thank you.
: CLAPACK seems to be designed in Linux? I don't
: know how to make it work in Windows.
: I found a book 'numeical recipes in C'. Maybe
: I will use SVD now.
|
h*******m
发帖数: 772 | 5 如果不是很大的矩阵的话，比如小于10000*10000的，
用svd比较合适。但对于更大型的稀疏矩阵方程，lsqr更
优越一些，其特点是占用内存比svd小很多，但其解的
可靠性也相应比svd小一些。
【在 b*****y 的大作中提到】
: You can use subroutines in CLAPACK.
: Dense LS-problem is usually solved using:
: 1. Normal equation
: 2. QR factorization
: 3. SVD
|
