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.
Scoring a black box model
From the course: Executive Guide to Predictive Modeling Strategy at Scale
Scoring a black box model
- [Instructor] What about scoring black box models? So called black box models are just models that have gotten sufficiently complex that it becomes extremely difficult to tell a story about your model. In other words you can't really interpret it very easily. Now, black box models are often very accurate. And they come in multiple forms. This artificial neural net is just one of many models that falls into this category. What people like about them is they can be incredibly accurate. But they have some disadvantages that we have to mention here. Now remember, each line in this diagram represents a coefficient. So they're getting quite complex. That also means that getting your black box model out of one platform and into another can be a challenge. So if you're building the model in once piece of software but then deploying it somewhere else you have to deal with that. One great way to do it is using something called predictive model markup language or PMML. Now this has been around…
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.