predict.mactivate_fit_gradient_logistic_01: Predict from Fitted Gradient Logistic Model

In mactivate: Multiplicative Activation

Description

Predict using fitted model returned by f_fit_gradient_logistic_01.

Usage

## S3 method for class 'mactivate_fit_gradient_logistic_01'
predict(object, X0, U0=NULL, mcols, ...)

Arguments

object	A list of class 'mactivate_fit_gradient_logistic_01' as returned by f_fit_gradient_logistic_01().
X0	Numeric matrix, N x d. Model ‘primary effect’ inputs.
U0	Numeric matrix with N rows. Inputs to pass to activation layer.
mcols	Scalar non-negative integer specifying which first columns of W to use.
...	Nothing else is required for this extension of the predict() function.

Details

If U0 is not provided, X0 will be passed to activation layer.

Value

A named list with 2 elements:

y0hat	Vector of length N. Linear predictions
p0hat	Vector of length N. Probability predictions. Similar to setting type='response' when predicting from glm logistic fitted model

See Also

f_fit_gradient_logistic_01.

Examples

####### Please see examples in the fitting functions

mactivate documentation built on Aug. 2, 2021, 5:07 p.m.