isIndep {dlsem} | R Documentation |
Conditional independence check
Description
Conditional independence between two variables is checked using the d-separation criterion (Pearl, 2000, page 16 and following).
Usage
isIndep(x, var1 = NULL, var2 = NULL, given = NULL, conf = 0.95, use.ns = FALSE)
Arguments
x |
An object of class |
var1 |
The name of the first variable. |
var2 |
The name of the second variable. |
given |
A vector containing the names of conditioning variables. If |
conf |
The confidence level for each edge: only edges with statistically significant causal effect at such confidence are considered. Default is 0.95. |
use.ns |
A logical value indicating whether edges without statistically significant causal effect (at level |
Value
Logical
Note
Conditional independence is checked statically, that is the whole history of conditioning variables is supposed to be known.
The result is unchanged if arguments var1
and var2
are switched.
Dependence is a necessary but not sufficient condition for causation: see the discussion in Pearl (2000).
References
J. Pearl (2000). Causality: Models, Reasoning, and Inference. Cambridge University Press. Cambridge, UK. ISBN: 978-0-521-89560-6
See Also
Examples
data(industry)
indus.code <- list(
Consum~ecq(Job,0,5),
Pollution~ecq(Job,1,8)+ecq(Consum,1,7)
)
indus.mod <- dlsem(indus.code,group="Region",exogenous=c("Population","GDP"),data=industry,
log=TRUE)
isIndep(indus.mod,"Job","Pollution",given=c("Consum"))