| Title: | Sigmoid Functions for Machine Learning |
| Version: | 1.4.0 |
| Description: | Several different sigmoid functions are implemented, including a wrapper function, SoftMax preprocessing and inverse functions. |
| Depends: | R (≥ 3.2.2) |
| Encoding: | UTF-8 |
| License: | GPL-3 |
| RoxygenNote: | 7.2.0 |
| Suggests: | covr, knitr, rmarkdown, ggplot2, testthat |
| VignetteBuilder: | knitr |
| NeedsCompilation: | no |
| Packaged: | 2022-06-18 14:22:15 UTC; bquast |
| Author: | Bastiaan Quast [aut, cre] |
| Maintainer: | Bastiaan Quast <bquast@gmail.com> |
| Repository: | CRAN |
| Date/Publication: | 2022-06-18 14:40:02 UTC |
Gompertz
Description
maps numeric vector using Gompertz function
Usage
Gompertz(x, a = 1, b = 1, c = 1)
Arguments
x |
input vector |
a |
see details |
b |
see details |
c |
see details |
SoftMax
Description
SoftMax preprocessing
Usage
SoftMax(x, lambda = 2)
Arguments
x |
input vector |
lambda |
see details |
SoftPlus
Description
maps numeric input vector using SoftPlus function
Usage
softplus(x)
Arguments
x |
input vector |
Inverse Gompertz
Description
maps numeric vector using Gompertz function
Usage
inverse_Gompertz(x)
Arguments
x |
input vector Gompertz values |
Leaky Rectified Linear Unit
Description
maps numeric vector using leaky ReLU function
Usage
leakyrelu(x)
Arguments
x |
input vector |
Standard Logistic
Description
maps numeric vector using logistic function
Usage
logistic(x, k = 1, x0 = 0)
Arguments
x |
input vector |
k |
see details |
x0 |
see details |
Logit
Description
maps numeric vector using logit function
Usage
logit(x)
Arguments
x |
input vector |
Rectified Linear Unit
Description
maps numeric vector using ReLU function
Usage
relu(x)
Arguments
x |
input vector |
ReLU Derivative
Description
Converts output of ReLU function to its derivative.
Usage
relu_output_to_derivative(x)
Arguments
x |
vector or ReLU values |
Sigmoid
Description
computes sigmoid nonlinearity
Usage
sigmoid(
x,
method = c("logistic", "Gompertz", "tanh", "ReLU", "leakyReLU"),
inverse = FALSE,
SoftMax = FALSE,
...
)
Arguments
x |
numeric vector |
method |
type of sigmoid function |
inverse |
use the inverse of the method (reverses) |
SoftMax |
use SoftMax preprocessing |
... |
arguments to pass on the method |
Examples
# create input vector
a <- seq(-10,10)
# use sigmoid with default standard logistic
( b <- sigmoid(a) )
# show shape
plot(b)
# inverse
hist( a - sigmoid(b, inverse=TRUE) )
# with SoftMax
( c <- sigmoid(a, SoftMax=TRUE) )
# show difference
hist(b-c)
Sigmoid Derivative
Description
Convert output of sigmoid function to its derivative.
Usage
sigmoid_output_to_derivative(x)
Arguments
x |
vector of sigmoid values |
SoftPlus Derivative
Description
Convert output of SoftPlus function to its derivative.
Usage
softplus_output_to_derivative(x)
Arguments
x |
vector of SoftPlus values |
Tanh Derivative
Description
Convert output of tanh function to its derivative.
Usage
tanh_output_to_derivative(x)
Arguments
x |
vector of tanh values |