rsa.runDf: Run RSA on a tidied data frame subset (by group or on DF with...

In benpeloquin7/rrrsa: Run Rational Speech Act Models of Pragmatic Inference

Description

Expects tidied data with three required (1-3) and two optional fields (4, 5)

Usage

rsa.runDf(data, quantityVarName, semanticsVarName, itemVarName,
  costsVarName = NA, priorsVarName = NA, depth = 1, alpha = 1,
  usePriorEveryRecurse = FALSE)

Arguments

data,	tidied data
quantityVarName,	entity name we're quantifying over (i.e. "stars" in Peloquin & Frank (2016))
semanticsVarName,	literal listener semantic values for RSA computations
itemVarName,	unique items were comparing, probaby words (i.e. "degrees" in Peloquin & Frank (2016))
costsVarName,	costs variable name
priorsVarName,	priors variable name
depth,	recursive depth parameter
alpha,	decision noise parameter level
usePriorEveryRecurse,	boolean incorporate priors during each full recursion

Details

1) quanityVarName :: entity name we're quantifying over

2) itemVarName :: unique items were considering,

3) semanticsVarName :: literal listener semantics

4) optional costsVarName :: costs

5) optional priorsVarName :: priors

rsa-ready, tidied data should have a unique semantic value for each quantity * item combination. (see formatting of peloquinFrank_2Alts data).

Value

data frame with posterior predictions 'preds' column appended

Examples

rsa.runDf(peloquinFrank_2Alts, "stars", "speaker.p", "words")

benpeloquin7/rrrsa documentation built on May 12, 2019, 2:08 p.m.