I am looking for an implementation based on numpyregular least-squares that would allow the fit to be updated with more observations. Something like strings algorithm applied statistics AS 274 or the R biglm.
numpy
biglm
Otherwise, the procedure for updating the QR decomposition with new lines will also be interesting.
Any pointers?
scikits.statsmodels has a recursive OLS that updates the reverse X'X in the sandbox, which can be used for this. (Used only to calculate OLS recursive residuals.)
Nathaniel Smith OLS, , scipy-user. X'X.
, econpy .
Pandas OLS, -.
Nathaniels biglm. , ( ).
, , , . python (-wrapped), QR.
: . http://mail.scipy.org/pipermail/scipy-dev/2010-February/013853.html
cholmod qr cholesky, , , Windows, , incremental_qr .
http://mail.scipy.org/pipermail/scipy-dev/2010-February/013844.html
pythonequations http://code.google.com/p/pythonequations/downloads/list, , , scipy numpy. http://zunzun.com - - ( ). . , - - , .
James Phillips 2548 Vera Cruz Drive Birmingham, AL 35235 USA zunzun@zunzun.com
, :
AFAIK, QR, , numpy, , .
QR
, x (of Ax= b) k, () ( , k )?
x
Ax= b
k
LSQ.F90
gfortran-4.4 -shared -fPIC -g -o lsq.so LSQ.F90
Python,
from ctypes import cdll lsq = cdll.LoadLibrary('./lsq.so')
, .