relevance-package {relevance}R Documentation

Calculate Relevance

Description

Calculates Relevance and Significance values for simple models and for many types of regression models

Details

The DESCRIPTION file:

Package: relevance
Type: Package
Title: Calculate Relevance
Version: 1.0
Date: 2020-08-26
Author: Werner A. Stahel
Maintainer: Werner A. Stahel <stahel@stat.math.ethz.ch>
Depends: stats
Suggests: survival, knitr
VignetteBuilder: knitr
Description: Calculates Relevance and Significance values for simple models and for many types of regression models
License: GPL-2
Repository: R-Forge
Repository/R-Forge/Project: regdevelop
Repository/R-Forge/Revision: 183
Repository/R-Forge/DateTimeStamp: 2020-09-23 20:40:36
Date/Publication: 2020-09-23 20:40:36

Index of help topics:

confintF                Confidence Interval for the Non-Central F and
                        Chisquare Distribution
dropNA                  Drop NA Values from Vector
getcoeftable            Extract Componente of a Fit
getopt                  Get Option that Influence the Output of
                        Relevance Functions
relevance-package       Calculate Relevance
rlstats                 Calculate Relevance and Significance Statistics
sumNA                   Count NA Values
termeffects             all coefficients of a model fit
termtable               Statistics for Linear Models, Including
                        Relevance Statistics

This package is being built.

Relevance is a measure that expresses the relevance of an effect. The simplest case is a single sample supposedly normally distributed observations, and we are interested in the expectation, estimated by the mean of the observations. There is a threshold for the expectation, below which an effect is judged too small to be of interest. The estimated relevance Rle is then simply the estimated effect divided by the threshold. If it is larger than 1, the effect is thus judged relevant. The two other values that characterize the relevance are the limits of the confidence interval for the relevance, called the secured relevance Rls and the potential relevance Rlp. If Rle $>1$, then one might say that the effect is "significantly relevant".

Another useful measure, meant to replace the p-value, is the "significance" Sg0. In the simple case, it divides the estimated effect by the critical value of the (t-) test statistic. Thus, the statistical test of the null hypothesis of zero expectation is significant if Sg0 is $>1$.

These measures are also calculated for the comparison of two groups, for proportions, and most importantly for regression models. For models with linear predictors, relevances are obtained for standardized coefficients as well as for the effect of dropping terms.

The most important functions are

rlstats

For models with a linear predictor, the results are given by tables. A coeftable is produced by rlstats. termtable contains the relevance of the terms as characterized by the effect of dropping the term from the model formula. It also includes the relevance of the coefficients as produced by rlstats for terms with a single degree of freedom. termeffects calculates the relevances for the coefficients related to each term and is thus only of interest for terms with more than one degree of freedom.

Author(s)

Werner A. Stahel

Maintainer: Werner A. Stahel <stahel@stat.math.ethz.ch>

References

Stahel, Werner A. (2020). Measuring Significance and Relevance instead of p-values. In preparation.

Examples

  data(swiss)
  rr <- lm(Fertility ~ . , data = swiss)
  rt <- termtable(rr)
  rt
  names(rt)
  data.frame(rt)

[Package relevance version 1.0 Index]