refers to a form of stepwise regression in which the least important variables are systematically removed from the prediction equation. Eventually, a preset criterion is reached.

BACKWARD ELIMINATION: "Backward elimination is a statistical procedure that removes the least important variables in a step-wise manner, leaving only the most important ones."
Cite this page: N., Pam M.S., "BACKWARD ELIMINATION," in, April 7, 2013, (accessed May 13, 2021).