continuousProbs.bayes: Conditional probability fitting step for Continuous data

In tbonza/supml: "Algorithms for Supervised Machine Learning"

Description

Only 'gaussian' kernel currently supported. Gaussian kernel requies mean mu given y in addition to the priors for y. The Gaussian kernel standard deviation is specified by hyperparameter lambda and not required for conditional probability computation. Priors for y are computed for the model rather than duplicated for each continuous feature j.

Usage

## S3 method for class 'bayes'
continuousProbs(object, X, y, continuous_cols)

Arguments

object	bayesian s3 object with mappings set
X	entire feature matrix
y	target vector, factor
continuous_cols	int vector of column positions

Details

This method should only be used for customizing your own Bayesian classifier.

Value

mean for each column j given a y class label; 'mu|j|y'

tbonza/supml documentation built on May 17, 2019, 3:14 a.m.