ci.mm: Symmetric conditional independence test with mixed data

In MXM: Feature Selection (Including Multiple Solutions) and Bayesian Networks

Symmetric conditional independence test with mixed data

Description

Symmetric conditional independence test with mixed data.

Usage

ci.mm(ind1, ind2, cs = NULL, dat, type, rob = FALSE, R = 1) 
ci.fast(ind1, ind2, cs = NULL, dat, type, rob = FALSE, R = 1)
ci.mm2(ind1, ind2, cs = NULL, suffStat) 
ci.fast2(ind1, ind2, cs = NULL, suffStat) 

Arguments

ind1	The index of the one variable to be considered.
ind2	The index of the other variable to be considered.
cs	The index or indices of the conditioning set of variable(s). If you have no variables set this equal to 0.
dat	A data.frame with numerical, binary, nominal and ordinal variables only.
type	This is obsolete basically, but we need it here, so that the functions ci.mm and ci.fast have the same signatures as in cat.ci, condi and dist.condi.
rob	This is obsolete basically, but we need it here, so that the functions ci.mm and ci.fast have the same signatures as in cat.ci, condi and dist.condi.
R	This is obsolete basically, but we need it here, so that the functions ci.mm and ci.fast have the same signatures as in cat.ci, condi and dist.condi.
suffStat	This is a list with only the dataset and the name must be "dataset". A data.frame with numerical, binary, nominal and ordinal variables only.

Details

The functions "ci.mm" and "ci.fast" are general functions to be used anywhere. The functions "ci.mm2" and "ci.fast2" are designed to be accepted by the command "pc" in the package "pcalg". The functions "ci.mm2" and "ci.fast2" can be fed in the "pc" function of the "pcalg" package in order to produce a PDAG with mixed data using the PC algorithm. For more information see the relevant paper in the references.

The "ci.mm" and "ci.fast" work with linear, logistic, multinomial and ordinal regression, whereas the "ci.mm2" and "ci.fast2" work with linear, logistic and ordinal regression only.

Value

A vector including the test statistic, it's associated p-value and the relevant degrees of freedom.

Author(s)

Michail Tsagris

R implementation and documentation: Michail Tsagris [e-mail]

References

Tsagris M. (2019). Bayesian network learning with the PC algorithm: an improved and correct variation. Applied Artificial Intelligence, 33(2): 101-123.

Tsagris M., Borboudakis G., Lagani V. and Tsamardinos I. (2018). Constraint-based Causal Discovery with Mixed Data. International Journal of Data Science and Analytics.

Spirtes P., Glymour C. and Scheines R. (2001). Causation, Prediction, and Search. The MIT Press, Cambridge, MA, USA, 3nd edition.

Sedgewick, A. J., Ramsey, J. D., Spirtes, P., Glymour, C., & Benos, P. V. (2017). Mixed Graphical Models for Causal Analysis of Multi-modal Variables. arXiv preprint arXiv:1704.02621.

See Also

pc.skel, mmhc.skel, cat.ci, condi

Examples

ci.mm(1, 2, dat = iris)
ci.mm(1, 5, dat = iris)
ci.fast(1, 5, dat = iris)
x <- iris
x[, 5] <- as.numeric(x[ ,5])  ## Caution:: this will be treated as ordered variable. 
x[, 5] <- factor(x[, 5], ordered = TRUE)
## ci.mm2 and ci.fast2 do not perform multinomial regression.
ci.mm2(1, 5, suffStat = list(dataset = x) )
ci.fast2(1, 5, suffStat = list(dataset = x) )

MXM documentation built on Aug. 25, 2022, 9:05 a.m.