estimate_control: Statistical Evaluation of Prediction Models (SEPM)

In maxwestphal/SEPM: Statistical Evaluation of Prediction Models

Statistical Evaluation of Prediction Models (SEPM)

Description

Control the estimation process.

Usage

estimate_control(method, type = c("generic", "probability"), mapping)

Arguments

method	character, name of the estimation method
type	either "generic" (default) or "probability", the latter should only be used for classification tasks
mapping	a function which specifies how the estimation is conducted, see details.

Details

If type="generic", mapping needs to be a function with input comparison and output list(theta.hat = ..., sigma.hat = ...) which is computed based on the input comparison matrix. If type="probability", mapping needs to to a function with input argument p (a probaility (matrix)) and output a modified (e.g. shrunk) probability with the same dimensions as p.

Value

A SEPM.estimate.control object which is passed to estimate.

maxwestphal/SEPM documentation built on April 12, 2024, 12:09 a.m.