compute_marginal_likelihood: Compute Marginal Likelihood

View source: R/gp_inference.R

compute_marginal_likelihoodR Documentation

Compute Marginal Likelihood

Description

compute_marginal_likelihood Computes Marginal likeilhood for the GPC Model, when using Expectation Prop. Used for as model evidence in order to compute optimal parameters/ hyperparameters for the model. Implemented so far only for the GaussKernel kernel function. Might be expanded in the future.

Usage

compute_marginal_likelihood(X, kernel_param, class_labels)

Arguments

X

(matrix) : The design matrix of (S/D) EC curves

kernel_param

(float) : Bandwidth parameter for the Gaussian Kernel

class_labels

(vector) : vector of +1,-1 of class labels.

Value

log_Z (float): the EP log likelihood given the parameters.


lcrawlab/SINATRA documentation built on Sept. 13, 2023, 2 p.m.