make_data_probabilities: Generates a probability distribution over possible data...

In lilymedina/gbiqqtools: Tools for generalized BIQQ

Description

NOTE: This needs to be checked for whether it is taking account of strategy probabilities properly

Usage

make_data_probabilities(model, pars, possible_data)

Arguments

model	A causal model as created by make_model
given	A data frame with observations
subset	data strategy

Value

A dataset

Examples

library(dplyr)
model <- make_model("X->M->Y")  %>%
   set_restrictions(causal_type_restrict = "Y[M=1]<Y[M=0] | M[X=1]<M[X=0] ") %>%
   set_parameter_matrix()

model <- set_parameters(model, type = "flat")
possible_data <- make_possible_data(model, N= 2)
make_data_probabilities(model, pars = model$parameters, possible_data)

given <- data.frame(X = c(0,0,0,1,1,1), M = NA, Y = c(0,0,1,0,1,1)) %>%
  collapse_data(model)
possible_data <- make_possible_data(model, given = given, condition = "X==1 & Y==1", vars = "M", within = TRUE )
make_data_probabilities(model, pars = model$parameters, possible_data)

lilymedina/gbiqqtools documentation built on Nov. 4, 2019, 4:32 p.m.