scikits.statsmodels is a pure python package that provides classes and functions for the estimation of several categories of statistical models. These currently include linear regression models, OLS, GLS, WLS and GLS with AR(p) errors, generalized linear models for six distribution families and M-estimators for robust linear models. An extensive list of result statistics are avalable for each estimation problem
License: Simplified BSD
Repository: http://code.launchpad.net/statsmodels
Pypi: http://pypi.python.org/pypi/scikits.statsmodels
Mailing List: http://groups.google.com/group/pystatsmodels?hl=en
Bug Tracker: https://bugs.launchpad.net/statsmodels
Installation:
easy_install scikits.statsmodels
or get the source from pypi, sourceforge, or from the launchpad repository and
setup.py install or, if this does not work, try
setup.py build install
Usage:
Get the data, run the estimation, and look at the results. For example, here is a minimal ordinary least squares case
import numpy as np
import scikits.statsmodels as sm
# get data
nsample = 100
x = np.linspace(0,10, 100)
X = sm.tools.add_constant(np.column_stack((x, x**2)))
beta = np.array([1, 0.1, 10])
y = np.dot(X, beta) + np.random.normal(size=nsample)
# run the regression
results = sm.OLS(y, X).fit()
# look at the results
print results.summary()
and look at `dir(results)` to see some of the results
that are available
Note: Due to our infrequent official releases, we want to point out that the trunk branch in the launchpad repository will have the most recent code and is usually stable and tested and fine for daily use.