01
Domain 01
The Prediction Lifecycle
Framing a prediction problem, beating the baseline, splitting honestly, and why accuracy lies.
6 questions
- Q01Which initiative is predictive analytics?Four analytics projects on the roadmap. Only one of them predicts.
- Q02Which framing is complete enough to start building a model?Unit, target, horizon, decision — the framing step everyone skips.
- Q03What accuracy does the do-nothing baseline (predict "no churn" for everyone) achieve, and what does that say about the vendor's model?A model beats "predict the majority" by two points. Is that a win?
- Q04What is wrong with the reported 84.1%?Train error falls, test error rises. Name the disease.
- Q05What is the diagnosis, and what is the sensible first move?Why the single most important line in the notebook is the split.
- Q06Which single change most directly fixes the problem?Stratify or not? A 1% fraud rate makes the answer obvious — once you see it.