Introduction to Uplift Modeling
summary
Introduces uplift modeling with concrete code and a reproducible notebook, shows how S-learner, T-learner and X-learner actually get trained and combined, and uses sklearn gradient-boost classifiers inside causalml/scikit-uplift style pipelines. Demonstrates simulated and real-data results, shows uplift curves and a percentile/cumulative gain evaluation that finds most lift around ~60–70% of targets, and discusses when unconfoundedness or sampling breaks the estimates.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 1:08Motivation
- 4:07Notation
- 7:23Data Collection
- 9:21Meta Learners
- 13:20X Learners
- 17:02Python implementations
- 18:27Evaluation
- 20:26Uplift by percentile
- 22:22Cumulative gain chart
- 22:51Cumulative sum
- 23:29Cumulative gain
- 25:33Uplift curve
- 30:08Random model
- 31:09In practice
- 33:44Notebook
- 35:25Recommendation
- 35:59Statistical Powers
- 36:17Questions
related talks
32mWhy start using uplift models for more efficient marketing campaigns
43mBeyond Just Prediction: Causal Thinking in Machine Learning
38mHow to Enhance Customer Targeting in Marketing
13mOptimizing Subscription Services with Uplift Modeling
37mCreating correct and capable classifiers
19mUnveiling the Potential: Exploring Upselling, pLTV, and ML for Operations
01h 00mAI Powered Uplift Modeling and Stream Forecasting in Music Industry
37mActive Teaching, Human Learning