Cross-entropy gradient for one class (morie.fn grn024)

Selects column k (0-based) of the grxeng gradient rather than recomputing it.

Usage

morie_geron_ch4_cross_entropy_gradient_vector(X, Y, Theta, k)

Arguments

X, Y, Theta

Design, labels and (n, K) parameters.

k

0-based class.

Value

List with gradient, class, gradient_norm, mean_error.

Examples

set.seed(3)
X <- cbind(1, rnorm(15))
Y <- diag(3)[sample(1:3, 15, replace = TRUE), ]
Theta <- matrix(0, 2, 3)
morie_geron_ch4_cross_entropy_gradient_vector(X, Y, Theta, k = 1)
#> $gradient
#> [1] -6.661338e-17  4.001049e-02
#> 
#> $class
#> [1] 1
#> 
#> $gradient_norm
#> [1] 0.04001049
#> 
#> $mean_error
#> [1] -3.70255e-17
#> 
#> $estimate
#> [1] -6.661338e-17  4.001049e-02
#> 
#> $n
#> [1] 15
#>