Package

weka.classifiers.bayes

Synopsis

Implements Bayesian Logistic Regression for both Gaussian and Laplace Priors.

For more information, see

Alexander Genkin, David D. Lewis, David Madigan (2004). Large-scale bayesian logistic regression for text categorization. URL http://www.stat.rutgers.edu/~madigan/PAPERS/shortFat-v3a.pdf.

Available in Weka 3.6.x - 3.7.1. Available via the package management system for Weka >= 3.7.2 (bayesianLogisticRegression).

The table below describes the options available for BayesianLogisticRegression.

Option	Description
debug	Turns on debugging mode.
hyperparameterRange	Hyperparameter value range. In case of CV-based Hyperparameters, you can specify the range in two ways: Comma-Separated: L: 3,5,6 (This will be a list of possible values.) Range: R:0.01-316,3.16 (This will take values from 0.01-316 (inclusive) in multiplications of 3.16
hyperparameterSelection	Select the type of Hyperparameter to be used.
hyperparameterValue	Specific hyperparameter value. Used when the hyperparameter selection method is set to specific value
maxIterations	The maximum number of iterations to perform.
normalizeData	Normalize the data.
numFolds	The number of folds to use for CV-based hyperparameter selection.
priorClass	The type of prior to be used.
threshold	Set the threshold for classifiction. The logistic function doesn't return a class label but an estimate of p(y=+1\|B, x(i)). These estimates need to be converted to binary class label predictions. values above the threshold are assigned class +1.
tolerance	This value decides the stopping criterion.

The table below describes the capabilites of BayesianLogisticRegression.

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