Package 'grnn'

GRNN

Description

General regression neural network.

Details

The program GRNN implements the algorithm proposed by Specht (1991).

Author(s)

Pierre-Olivier Chasset

References

Specht D.F. (1991). A general regression neural network. IEEE Transactions on Neural Networks, 2(6):568-576.

---

Guess

Description

Infers the value of a new observation.

Usage

guess(nn, X)

Arguments

nn	A trained and smoothed General regression neural network.
X	A vector describing a new observation.

See Also

grnn-package

Examples

n <- 100
set.seed(1)
x <- runif(n, -2, 2)
y0 <- x^3
epsilon <- rnorm(n, 0, .1)
y <- y0 + epsilon
grnn <- learn(data.frame(y,x))
grnn <- smooth(grnn, sigma=0.1)
guess(grnn, -2)
guess(grnn, -1)
guess(grnn, -0.2)
guess(grnn, -0.1)
guess(grnn, 0)
guess(grnn, 0.1)
guess(grnn, 0.2)
guess(grnn, 1)
guess(grnn, 2)

---

Learn

Description

Create or update a General regression neural network.

Usage

learn(set, nn, variable.column = 1)

Arguments

set	Data frame representing the training set. The first column is used to define the category of each observation (set category.column if it is not the case).
nn	A General regression neural network with or without training.
variable.column	The field number of the variable (1 by default).

See Also

grnn-package

---

Smooth

Description

Smooth a General regression neural network.

Usage

smooth(nn, sigma)

Arguments

nn	A trained General regression neural network.
sigma	A scalar.

See Also

grnn-package

---