regout

Among the considerations in the use of analysis of regression, outliers or bad values can seriously disturb the least-squares fit. They falls far from the line implied by the rest of the data. If these points are really outliers, then the estimate of the intercept may be incorrect and the residual mean square may be an inflated estimate of the true variance. There are methods for scaled residuals useful in finding observations that are outliers. One of them is the externally studentized residual, usually called R-student. It is based on the fact that the MSRes is an internally generated estimate of the variance obtained from fitting the model to all n observations and it is necessary to get an estimation based on a data set with the ith observation removed. This statistic follows a Student t-distribution. But one could use a Bonferroni-type approach and compare all n values of t_i to t_(alpha/2*n),n-p-1 to provide guidance regarding outliers.

Inputs:
D - matrix data (=[X Y]) (last column must be the Y-dependent variable). (X-independent variable entry can be for a simple [X], multiple [X1,X2,X3,...Xp] or polynomial [X,X^2,X^3,...,X^p] regression model).
alpha - significance (default = 0.05).

Outputs:
A complete summary (table and/or plot) of the outliers diagnostic test.