Learn R Programming

evidence (version 0.8.10)

Analysis of Scientific Evidence Using Bayesian and Likelihood Methods

Description

Bayesian (and some likelihoodist) functions as alternatives to hypothesis-testing functions in R base using a user interface patterned after those of R's hypothesis testing functions. See McElreath (2016, ISBN: 978-1-4822-5344-3), Gelman and Hill (2007, ISBN: 0-521-68689-X) (new edition in preparation) and Albert (2009, ISBN: 978-0-387-71384-7) for good introductions to Bayesian analysis and Pawitan (2002, ISBN: 0-19-850765-8) for the Likelihood approach. The functions in the package also make extensive use of graphical displays for data exploration and model comparison.

Copy Link

Version

Install

install.packages('evidence')

Monthly Downloads

296

Version

0.8.10

License

GPL (>= 2)

Maintainer

Robert Hulst

Last Published

May 15th, 2018

Functions in evidence (0.8.10)

Bayesian analysis of a Normal sample using a SIR prior.

A contingency table for heart attacks and aspirin use.

Bayesian analysis of one sample from a Normal distribution with imprecise priors.

A contingency table of 16 British youths categorized as juvenile delinquents or not, and as wearing glasses or not.

Human body fat and several covariates for calculating it.

Morphology of horseshoe crabs.

Likelihood analysis of the binomial parameter for one sample.

generates the 100 * (1 - alpha)% most probable interval of a distribution of empirical values

computes the Negative Predictive Value.

calculates the positive predictive value (PPV) of a diagnostic test.

Bayesian analysis of n >= 2 Normal means with standard improper reference priors.

A dot plot is produced for several related models showing for each model its LOOIC-value with its credible interval.

Plots a simple strip chart of the observations with group means and grand mean.

Bayesian regression model comparison with Bayes factors.

A support function that calculates the sum of squares of a data vector.

produces a Normality plot for the argument surrounded by eight other Normality plots for Normal distributions having the same mean and standard deviation as the argument

overdispersionCheck

A robust comparison of the location and the scale of the input vector.

Bayesian analysis of the binomial parameters for two samples.

Data of the growth of tissue cultures on five different media.

Data on lung capacity of 654 children and adolescents.

Simon Newcomb's measurements of the speed of light

simulates Bayesian updating of the binomial parameter \(\pi\).

Bayesian analysis of a 2 x 2 contingency table.

Carcinogenesis data on rats exposed to a carcinogen.

Bayesian t-test using reference priors.

Bayesian analysis of the means of two Normal samples using SIR priors.

Contingency Table Analysis in different ways

Universal Fisherian significance test with confidence interval.

Data on body temperature, heart rate, and gender of 130 human subjects.

weight gain in rats

function to plot diverse Beta distributions for use as Binomial priors

a simple example of the bias--variance trade-off.

Simulated clutch size data for birds with different nesting locations.

Mortality due to heart infarct in smokers and non-smokers.

Data on the incidence of hypertension and three indicator variables.

Mortality data of moth larvae due to increasing doses of insecticide.

Made-up data that are not unlike the actual data collected by Nespolo et al.(2003).

Plotting routine for dataframes of looic values.

Likelihood analysis of the binomial parameters for two samples.

Conversion of 2 props input to 2x2 contingency table

Computes the posterior probability of having a certain disease from prevalence, sensitivity, and specificity data.

Charles Darwin's (1876) data on the fecundity of selfed and crossed corn plants.

Conversion of a frequentist p-value to the lower bound of the Bayes factor against the null hypothesis assuming equal odds of the null and the alternative

Conversion of a frequentist p-value to a lower bound of the posterior probability that the null hypothesis is true assuming equal odds of the null and the alternative

evidence-package

evidence: Functions and Data for Bayesian and Likelihood Analysis

Made-up data for a balanced one-way anova.

Made-up data for an unbalanced one-way anova.

Made-up data to illustrate Simpson's paradox.

Bayesian analysis of the binomial parameter for one sample.