Cost function for CDEN model fitting

Description

The maximum likelihood cost function used for CDEN model fitting. Calculates the negative of the logarithm of the likelihood. A normal distribution prior can be set for the magnitude of the input-hidden layer weights, thus leading to weight penalty regularization.

Usage

1
2
cadence.cost(weights, x, y, n.hidden, hidden.fcn, distribution, sd.norm,
             valid)

Arguments

weights

weight vector of length returned by cadence.initialize.

x

matrix with number of rows equal to the number of samples and number of columns equal to the number of predictor variables.

y

column matrix of predictand values with number of rows equal to the number of samples.

n.hidden

number of hidden nodes in the CDEN model.

hidden.fcn

hidden layer transfer function.

distribution

a list that describes the probability density function associated with the predictand.

sd.norm

sd parameter for normal distribution prior for the magnitude of input-hidden layer weights; equivalent to weight penalty regularization.

valid

valid logical vector indicating which weights are non-zero or fixed at zero, i.e., due to use of parameters.fixed in distribution.

See Also

cadence.fit, optim, rprop