Level: Intermediate
Karen Copeland, Statistician, Boulder Statistics
There are many steps to building predictive models. One key step is identifying variables to include in your model. This is particularly challenging when you have an abundance of variables to choose from, many of which are likely not important. Thus, you have needles hiding in a haystack, how can you find the needles? I explore a variable selection process that includes predictor screening followed by generalized regression with lasso fitting followed by one-click bootstrapping.