Glossary · Phase 3 · Machine learning
Holdout group
Customers or data deliberately left out, so you can compare against them and see the real effect.
Try it now
A good model isn't the goal. A campaign that makes money is. Try lowering precision, or raising the cost per contact, and find where it stops paying. The figures here are examples to play with.
Learn it properly
Capstone: churn prediction · Predict which customers will leave, and measure it properly. About 45 minutes.
Other terms from the same lesson
- lift: How many times better a targeted group does than average: if 40% of the top 5 churn against 20% overall, the lift is 2.
- permutation importance: How much a model's score drops when one column is shuffled. A big drop means the model leans on that column.
- uplift modeling: Predicting who will change their behavior because you contacted them, not just who is likely to churn anyway.
Never trained an AI?
Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account.