RegBest function

Select variables in multiple linear regression

Find an optimal submodel

RegBest(y,x, int = TRUE, wt=NULL, na.action = na.omit,
    method=c("r2","Cp", "adjr2"), nbest=1)

Arguments

• y: A response vector
• x: A matrix of predictors
• int: Add an intercept to the model
• wt: Optional weight vector
• na.action: Handling missing values
• method: Calculate R-squared, adjusted R-squared or Cp to select the model. By default a the F-test on the r-square is used
• nbest: number of best models for each set of explained variables (by default 1)

Returns

Returns the objects - all: gives all the nbest best models for a given number of variables

• best: the best model

Author(s)

Francois Husson francois.husson@institut-agro.fr

See Also

lm

Examples

data(milk)
res = RegBest(y=milk[,6],x=milk[,-6])
res$best