current community

your communities

more stack exchange communities

Stack Exchange
tour help
stack overflow careers
Take the 2-minute tour ×
Stack Overflow is a question and answer site for professional and enthusiast programmers. It's 100% free, no registration required.
up vote 1 down vote favorite
2

Is it possible to perform multi-variate regression in Python using NumPy?

The documentation here suggests that it is, but I cannot find any more details on the topic.

share|improve this question

3 Answers 3

up vote 3 down vote accepted

Yes, download this ( http://www.scipy.org/Cookbook/OLS?action=AttachFile&do=get&target=ols.0.2.py ) from http://www.scipy.org/Cookbook/OLS

Or you can install R and a python-R link. R can do anything.

share|improve this answer
    
I'm currently using R, but I was considering making the calculation just python, for ease of sharing. –  celenius May 18 '10 at 20:47
    
It doesn't work any longer (I changed it to conform to Python 3, so it run; but the very first test crashed Python interpreter with no error message). Anyone knows how to update it? –  max Dec 15 '11 at 20:35
up vote 2 down vote

The webpage that you linked to mentions numpy.linalg.lstsq to find the vector x which minimizes |b - Ax|. Here is a little example of how it can be used:

First we setup some "random" data:

import numpy as np
c1,c2 = 5.0,2.0
x = np.arange(1,11)/10.0
y = c1*np.exp(-x)+c2*x
b = y + 0.01*max(y)*np.random.randn(len(y))
A = np.column_stack((np.exp(-x),x))
c,resid,rank,sigma = np.linalg.lstsq(A,b)
print(c)
# [ 4.96579654  2.03913202]
share|improve this answer
up vote 1 down vote

You might want to look into the scipy.optimize.leastsq function. It's rather complicated but I seem to remember that being the thing I would look to when I wanted to do a multivariate regression. (It's been a while so I could be misremembering)

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.