Run CMST tests the causal direction between the first outcome (y1)
and all other outcomes (y2, ...)
,
adjusting for the driver and covariates. The models tested (ignoring covariates and noise) are:
M1.y2.y1.z
: y2 <- y1 <- z
M2.y1.y2.z
: y1 <- y2 <- z
M3.y1.z.y2
: y1 <- z -> y2
M4.y12.z
: y1 <- z -> y2
and y1 <-> y2
in which the arrows indicate direction of causality. For M4
, the directionality between y1
and y2
is ambiguous.
The noise on outcomes yi
is assumed to be normal.
The driver z
is assumed to be causal for one or more outcomes (y1, y2, ...)
, and may be categorical or continuous.
The covariates X
, qualitative or quantitative, may act additively or be interactive with the driver.
The driver
is the driver for outcomes, with covariates acting on outcomess.
The resp_names
, addcov
and intcov
names must all be valid names in the outcomes
data frame.
1 2 3 |
driver |
data frame with drivers |
outcomes |
data frame with outcomes |
addcov, intcov |
additive and interactive covariate names for responses |
method |
method for CMST test (parametric, non-parametric, joint or all three); can provide more than one value. |
penalty |
type of information criteria penalty (for BIC or AIC) |
verbose |
verbose output if |
fitFunction |
log likelihood calculation function; see |
... |
possible additional arguments |
resp_names |
response names |
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.