get_prob: Computes model class conditional probabilities

In IshmaelBelghazi/gpuClassifieR: gpuClassifieR C/CUDA linear classifier for R

Description

Computes model class conditional probabilities

Usage

get_prob(object, feats, normalize = TRUE, log_domain = FALSE,
  backend = "R", ...)

## S3 method for class 'model.spec'
get_prob(object, feats, normalize = TRUE,
  log_domain = FALSE, backend = "R", ...)

Arguments

object	Linear classifier specification object.
feats	Numeric Matrix of features. Follows the usual convention of having one example per row. For a model with M features and dataset with N examples the matrix should be N \times K
normalize	Logical scalar. If TRUE return normalized probabilities
log_domain	Logical scalar. If TRUE returns log probabilities.
backend	Computation back-end ('R', 'C', or 'CUDA')
...	other arguments passed to specific methods

Value

Numeric Matrix. Class conditional probabilities matrix. Each row corresponds to the probabilities of one example. For model with K classes and a dataset of N examples the returned matrix should be N \times K

Methods (by class)

• model.spec: Computes model class conditional probabilities for linear classifier specification objects

Author(s)

Mohamed Ishmael Diwan Belghazi

Examples

# Generate random initial weights
w_init <- matrix(rnorm(784 * 10), 784, 10)
# construct model
linear_classifier <- Classifier(weights_init=w_init)
# Fetch training variables
feats <- mini_mnist$train$images
targets <- mini_mnist$train$labels
# Set decay coefficient
decay <- 0.01
# compute log probabilities of the training set using the three back-ends
log_prob_R <- get_prob(linear_classifier, feats, TRUE, TRUE, 'R')
log_prob_C <- get_prob(linear_classifier, feats, TRUE, TRUE, 'C')
log_prob_CUDA <- get_prob(linear_classifier, feats, TRUE, TRUE, 'CUDA')

IshmaelBelghazi/gpuClassifieR documentation built on May 7, 2019, 6:45 a.m.