VanillaBayes: vanillaBayes

In dlandy/WarpedBayes: Implements Design Pattern Discussed in Landy, Crawford, & Corbin

Description

Performs vanilla bayes (single category) on a set of stimuli.

Usage

vanillaBayes(stimuli, kappa = 0, tauStimuli = 1, tauCategory = 1,
  responses = NULL, mode = "prediction", responseGrid = c(0))

Arguments

stimuli	a vector of stimuli, between -inf and inf
kappa	The location of the category
tauStimuli	The precision of the stimulus traces: may be a single number or a vector
tauCategory	The precision of the category distribution
responses	an optional vector of responses. Should only be given if mode is "subjectiveLogLikelihood"
mode	What aspect should the function calculate? Legel choices include "prediction", "simulation", and "subjectiveLogLikelihood"

Details

This function assumes that the data are in a metric space (-inf, inf), with a single normally distributed generating category (with mean kappa and precision tauCategory). It further assumes a set of stimuli, which are normal distributions with means at the value of stimuli, and precision tauStimuli. It returns either the mean expected location of response to the stimuli (given the parameters), or if a set of responses is given, the log likelihood of the responses given the model.

If the kappa, tauStimuli, and tauCategory items are all more than length 1, and are length 2 less than the number of bins, then we pad them by negative and positive infinity.

Value

A vector containing mean predicted stimulus locations, or the log likelihood of the responses given the model

See Also

vanillaBayes

Examples

(0:1000/1000) %>% vanillaBayes(kappa=5) 
    # The  Bayesian normal-normal model typical to many analyses
(0:1000/1000) %>% psiLogOdds() %>% 
    vanillaBayes(kappa=5) %>% 
    psiLogOddsInverse()  #Gonzales & Wu, 1996
1:1000 %>% psiLog() %>% vanillaBayes() %>% psiLogInverse()  # Stevens Power Law
plot(-99:100/100, (-99:100/100) %>% multiCycle(references= c(-10, 0, 10)) %>% 
    psiLogOdds() %>% vanillaBayes(kappa=c(-1, 1), tauStimuli=10) %>% 
    psiLogOddsInverse() %>% 
    multiCycleInverse(references=c(-10, 0, 10))-(-99:100/100), 
        ylab="bias", xlab="stimulus");abline(0,0)

dlandy/WarpedBayes documentation built on May 29, 2019, 2:49 p.m.