Identifiers are functions that take as input a covariance matrix Sigma
corresponding to some latent digraph G and, from that covariance matrix,
identify some subset of the coefficients coresponding to the direct causal effects
in the latent digraph G. This function
takes as input the digraph G and creates an identifier
that does not identify any of the direct causal effects. This is useful as a
base case when building more complex identification functions.
createLFIdentifierBaseCase(graph)a function that takes as input a covariance matrix compatible with
the latent digraph defined by L and returns a list with two
named components:
Lambdaa matrix equal to the observed part of graph$L() but with NA values
instead of 1s
Omegaa matrix equal to graph$O() but with NA values for coefficients
not equal to zero.
When building more complex identifiers these NAs will be replaced
by the value that can be identified from the covariance matrix corresponding to G.
a LatentDigraph object representing
the latent-factor graph. All latent nodes in this graph should be
source nodes (i.e. have no parents).