From the course: Executive Guide to Predictive Modeling Strategy at Scale
Join today to access over 22,600 courses taught by industry experts or purchase this course individually.
Feature engineering
From the course: Executive Guide to Predictive Modeling Strategy at Scale
Feature engineering
- [Instructor] Pretty much everything we've been talking about in regards to data preparation is a kind of feature engineering. Now, the cross-industry standard process for data mining likes to call this data construction but the more common term these days is feature engineering. What it basically is is extracting the meaning out of variables by using a whole bunch of different formulas, so we've already seen that we might calculate a ratio or a delta comparing recent behavior to baseline behavior. We generate all kinds of variables through aggregating and restructuring. But once it's already in its basic form, one row per customer, or one row per patient, one row per loan, we're still going to continue doing formulas. Let me show you a couple of quick examples. Dates, for instance, are not very useful to these predictive models but the gaps between dates is very interesting to modeling, so we're going to be doing…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.