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.

Slow algorithms: Brute force

Slow algorithms: Brute force

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

Start my 1-month free trial

Slow algorithms: Brute force

“

- [Narrator] Now we're going to talk about three reasons why some algorithmic approaches are much slower than others. The first is that some modeling algorithms they tend to be of the machine learning style, perform brute-force calculations. Let me show you what I mean. This is tiny little data set, but we're trying to predict whether or not an expected mom is going to have a low birth weight baby. If we use a statistically based approach, what it's going to do is only perform about a dozen calculations or so between a dozen and two dozen. Specifically what it's going to do to figure out that that cut point between lower risk and higher risk is 107 pounds is it's going to break all of the expectant moms into deciles, less than 100 pounds going all the way up with the highest one being over 171 pounds. Then it's going to perform a calculation on just those 10 groups to decide where that cut point should be. If we're…

Contents