Stepaic Forward, This function performs Backward stepwise is generally better because starting with the full model has the advantage of considering the effects of all variables Details step uses add1 and drop1 repeatedly; it will work for any method for which they work, and that is determined by having a valid method for extractAIC. The function has been changed recently to allow parallel computation. interaction SL. g. Stepwise Model Selection Description This function is a front end to the stepAIC function in the MASS package. There is an "anova" component corresponding to the steps taken the mode of stepwise search, can be one of "both", "backward", or "forward", with a default of "both". Length) However, as mentioned by @BenBolker you should post a reproducible example with 19 In R stepwise forward regression, I specify a minimal model and a set of variables to add (or not to add): python science data backward regression variable feature-selection automated feature forward elimination stepwise-regression backward-elimination forward I am using the stepAIC function in R to do a bi-directional (forward and backward) stepwise regression. I do not understand what each return value from the function means. stepAIC () has a direction argument that can be set to 'backward' I'm implementing a logistic regression model in R and I have 80 variables to chose from. Often this procedure converges to a subset of features.