Package mekano :: Package ml :: Module logreg :: Class LogisticRegressionClassifier

Class LogisticRegressionClassifier

object --+
         |
        LogisticRegressionClassifier

A logistic regression classifier trained using CG. Supports prior mean on 'w'.

>>> lr = LogisticRegressionClassifier()
>>> lr.maxiter = 100
>>> lr.epsilon = 1e-8
>>> lr.c = 1.0             # How much more to weigh the +ve examples than -ve examples.
>>> lr.train(ds)           # see help for train()
>>> s = lr.score(av)

Instance Methods

[hide private]

__init__(...)
Initialize the LR classifier.

a new object with type S, a subtype of T

__new__(T, S, ...)

__repr__(x)
repr(x)

score(...)

train(...)
Train the model.

Inherited from object: __delattr__, __format__, __getattribute__, __hash__, __reduce__, __reduce_ex__, __setattr__, __sizeof__, __str__, __subclasshook__

Properties

[hide private]

LAMBDA

epsilon

maxiter

Inherited from object: __class__

Method Details

[hide private]

init(...)
(Constructor)

Initialize the LR classifier.

        mu can be a double value, or a list of values (at least as many as #features+1)
lambda is a double value that controls the strength of the prior.

Overrides: object.__init__

new(T, S, ...)

Returns: a new object with type S, a subtype of T
Overrides: object.__new__

repr(x)
(Representation operator)

repr(x)

Overrides: object.__repr__

train(...)

Train the model.

Parameters:

ds - A Dataset that contains docs as a list of AtomVectors and labels as a list of true/false objects.

Class LogisticRegressionClassifier

__init__(...) (Constructor)

__new__(T, S, ...)

__repr__(x) (Representation operator)

train(...)

init(...)
(Constructor)

new(T, S, ...)

repr(x)
(Representation operator)