SGD

Package

weka.classifiers.functions

Synopsis

COMING IN WEKA 3.7.2

Implements stochastic gradient descent for learning various linear models (binary class SVM, binary class logistic regression and linear regression). globally replaces all missing values and transforms nominal attributes into binary ones. It also normalizes all attributes, so the coefficients in the output are based on the normalized data. This implementation can be trained incrementally on (potentially) infinite data streams.

Options

The table below describes the options available for SGD.

Option	Description
debug	If set to true, classifier may output additional info to the console.
dontNormalize	Turn normalization off
dontReplaceMissing	Turn off global replacement of missing values
epochs	The number of epochs to perform (batch learning). The total number of iterations is epochs * num instances.
lambda	The regularization constant. (default = 0.0001)
learningRate	The learning rate. If normalization is turned off (as it is automatically for streaming data), thenthe default learning rate will need to be reduced (try 0.0001).
lossFunction	The loss function to use. Hinge loss (SVM), log loss (logistic regression) or squared loss (regression).
seed	The random number seed to be used.

Capabilities

The table below describes the capabilities of SGD.

Capability	Supported
Class	Missing class values, Binary class, Numeric class
Attributes	Binary attributes, Nominal attributes, Unary attributes, Numeric attributes, Empty nominal attributes, Missing values
Min # of instances	0