pdk.Math.UniVariate.Regression ($Date: 2002/12/04 10:16:11 $)
index
pdk/Math/UniVariate/Regression.py

Univariate statistical routines.

FOG 11.1999

 
Classes
            
_LinearRegressor
 
class _LinearRegressor
       
   Methods defined here:
__call__(self, X, Y, calcResiduals=1, calcR=1, calcVIF=1)
Purpose:(multiple) linear regression
Detail:

regress the vector/matrix of independent variables X on the independent variable Y using a SVD of X. Returns a dictionary with the following keys:

  • "coefficients": the vector of parameter estimates
  • "residuals": the residuals, if calcResicuals is set, or None
  • "R": the coefficient of multiple determination, if calcR is set, or None
  • "VIF": the Variance Inflation Factor(s), if calcVIF is set, or None.

Note that no precautions are taken against rank-deficient input. See Chambers(1977): Methods for Computational Data Analysis, Wiley, pp. 101 ff.

 
Data
             linear_regression = instance of _LinearRegressor
 
Author
            
$Author: gathmann $