From the course: Executive Guide to Predictive Modeling Strategy at Scale

Unlock the full course today

Join today to access over 22,600 courses taught by industry experts or purchase this course individually.

Modeling with missing data

Modeling with missing data

From the course: Executive Guide to Predictive Modeling Strategy at Scale

Start my 1-month free trial

Modeling with missing data

“

- [Instructor] Now we're going to talk about a topic that could easily warrant its own course, Modeling with Missing Data. But we're going to try to stay laser focused on the topic at hand, which is the implications of the decisions that we make throughout the entire process with a particular focus on scoring and deployment. You've got to be deployment focused. 100% of the times there's going to be some missing data. Then what happens in most projects is the modeler has to fix it and clean it so that it's not missing. But then at scoring, it's missing again. Everybody has to work together to ensure that the model will work at scoring. Modelers, especially if they're statisticians, are trained to repair and address data issues. This is a critical skill to have on the team, but don't forget to ask yourself if the data even belongs in the model. For instance, before the mortgage crisis, a lot of small real estate investors…

Contents