cvq2-class: Class '"cvq2"'

In cvq2: Calculate the predictive squared correlation coefficient

Description

The class "cvq2" extends class "q2" and is used to store information about the model calibration, its prediction performance and the cross-validation applied to determine the prediction performance.

Objects from the Class

Objects can be created by calls of the form new("cvq2", ...).

Slots

result

Contains three lists (fit, pred, cv) regarding the results from linear regression (model calibration, fit) and cross-validation (prediction power, pred and cv) for the given model.

output

A list of parameters like number formats, output restrictions or output targets.

Linear regression and prediction result list

These lists are inherted from the parent class q2. Differences caused by cross-validation appear in the prediction result list for:

data

For each observation the model parameters used for the prediction are stored additionally as well as the arithmetic mean of the training set

nTrainingSet

The number of elements in one training set (N-k) plus an eventually variation.

nTestSet

The number of elements in one test set (k) minus an eventually variation.

Cross-validation result list

testSetSizeVaries

True, if some test sets consist of k-1 elements.

nFold

modelData is randomly split into n equal sized (according to testSetSizeVaries) test sets for each individual run.

nRun

The number of runs each value is predicted.

Extends

Class "q2", directly.

Methods

show

Returns a comprehensive overview about the model calibration and the prediction performance.

Author(s)

Torsten Thalheim <[e-mail]>

Examples

  showClass("cvq2")

cvq2 documentation built on May 2, 2019, 8:29 a.m.