pyvision.ml.regression

Module regression

Created on Apr 27, 2012

Author: bolme

Classes

[hide private]

pdf2nll
This is a wraper class that converts a pdf to a negitive log likelihood function.

LogisticRegression
This object implements a logistic regression model.

Functions

[hide private]

logit_approx(p)

source code

maxLikelihoodEstimate(obs, params, pdf=None, nll=None)
Produces a maximum likelihood estimate of the parameters at least one pdf or nll needs to be specified.

source code

Variables

[hide private]

__package__ = 'pyvision.ml'

Function Details

[hide private]

maxLikelihoodEstimate(obs, params, pdf=None, nll=None)

source code

Produces a maximum likelihood estimate of the parameters at least one pdf or nll needs to be specified.

Parameters:

obs (list of numbers) - a list of numerical values that are the observations.
params (list of numbers) - a list or numpy array containing estimates of the parameters.
pdf (pdf(params,x) -> probablity density) - a function returning the probability density for a distribution.
nll (nll(params,obs) -> negitive log likelihood.) - a function returning the negitive log likelihood given the parameters and observations as arguments.