semfindr (version 0.1.9)

mahalanobis_predictors: Mahalanobis Distance On Observed Predictors

Description

Gets a lavaan_rerun() or lavaan::lavaan() output and computes the Mahalanobis distance for each case using only the observed predictors.

Usage

mahalanobis_predictors(
  fit,
  emNorm_arg = list(estimate.worst = FALSE, criterion = 1e-06)
)

Value

A md_semfindr-class object, which is a one-column matrix (a column vector) of the Mahalanobis distance for each case. The number of rows equals to the number of cases in the data stored in the fit object. A print method is available for user-friendly output.

Arguments

fit

It can be the output from lavaan, such as lavaan::cfa() and lavaan::sem(), or the output from lavaan_rerun().

emNorm_arg

No longer used. Kept for backward compatibility.

Author

Shu Fai Cheung https://orcid.org/0000-0002-9871-9448.

Details

For each case, mahalanobis_predictors() computes the Mahalanobis distance of each case on the observed predictors.

If there are no missing values, stats::mahalanobis() will be used to compute the Mahalanobis distance.

If there are missing values on the observed predictors, the means and variance-covariance matrices will be estimated by maximum likelihood using lavaan::lavCor(). The estimates will be passed to modi::MDmiss() to compute the Mahalanobis distance.

Supports both single-group and multiple-group models. For multiple-group models, the Mahalanobis distance for each case is computed using the means and covariance matrix of the group this case belongs to. (Support for multiple-group models available in 0.1.4.8 and later version).

References

Béguin, C., & Hulliger, B. (2004). Multivariate outlier detection in incomplete survey data: The epidemic algorithm and transformed rank correlations. Journal of the Royal Statistical Society: Series A (Statistics in Society), 167(2), 275-294.

Mahalanobis, P. C. (1936). On the generalized distance in statistics. Proceedings of the National Institute of Science of India, 2, 49-55.

Schafer, J.L. (1997) Analysis of incomplete multivariate data. Chapman & Hall/CRC Press.

Examples

library(lavaan)
dat <- pa_dat
# For illustration, select only the first 50 cases.
dat <- dat[1:50, ]
# The model
mod <-
"
m1 ~ a1 * iv1 +  a2 * iv2
dv ~ b * m1
a1b := a1 * b
a2b := a2 * b
"
# Fit the model
fit <- lavaan::sem(mod, dat)
summary(fit)

md_predictors <- mahalanobis_predictors(fit)
md_predictors