reglev: Identify good and bad leverage points and regression outliers

In abnormally-distributed/cvreg: Cross Validation and Robust Estimation Utilities

Description

Search for good and bad leverage points using the method advocated by Rousseuw and van Zomeren (1990). This supports regression on numeric outcomes assumed to follow a Gaussian-like distribution (i.e., not generalized linear models). Several options for calculating the robust mahalanobis distances and regression residuals are supported.

Usage

reglev(
  formula,
  data,
  plot = TRUE,
  interact = F,
  cov.method = c("mvv", "mcd", "mgv", "stu"),
  reg.method = c("lqd", "lta", "ts", "lad", "ols")
)

Arguments

formula	formula
data	data frame
plot	should it be plotted? defaults to TRUE.
interact	if TRUE you can click on points to identify which observation number to which a given point corresponds.
cov.method	The options are as follows: - "mvv" - Minimum Variance Vector. The default option. - "mcd" - Minimum (Regularized) Covariance Determinant - "mgv" - Minimum Generalized Variance - "stu" - Student's T Covariance Matrix
reg.method	The options are as follows: - "lqd" - Least Quantile Differences Regression GS-Estimator - "lta" - Least Trimmed Absolute Residuals Regression S-Estimator - "ts" - Theil-Sen estimator - "lad" - Least Absolute Deviations Regression Estimator - "ols" - Ordinary Least Squares Non-Robust Regression Estimator

Value

a plot or a list.

References

Rousseuw, P.J.; van Zomeren, B.C. (1990) Unmasking Multivariate Outliers and Leverage Points. Journal of the American Statistical Association. 85(411):633-639 doi:10.1080/01621459.1990.10474920

abnormally-distributed/cvreg documentation built on May 3, 2020, 3:45 p.m.