User Guide

Warning

Empulse is still under active development, and we make no promises about the stability of the API. Pin a version if you are worried about breaking changes!

• 1. Metrics
  • 1.1. Choosing the right metric
  • 1.2. Define your own cost-sensitive or value metric
• 2. Cost-sensitive and Value-driven models
  • 2.1. Cost-Sensitive Logistic Regression (CSLogit)
  • 2.2. Cost-Sensitive Gradient Boosting (CSBoost & B2Boost)
  • 2.3. Boosting Algorithms Custom Cost Functions
  • 2.4. Robust Cost-Sensitive Classification (RobustCS)
  • 2.5. Profit-Driven Logistic Regression (ProfLogit)
• 3. Model-Independent Preprocessing
  • 3.1. Bias Mitigation
  • 3.2. Cost-Proportionate Sampling
• 4. Cross-Validation with Instance-dependent Costs
  • 4.1. Enabling Metadata Routing
  • 4.2. What is Metadata Routing
  • 4.3. Gridsearch Example
• 5. Datasets
  • 5.1. Bank Telemarketing Upsell Campaign
  • 5.2. Churn in a TV Subscription Company
  • 5.3. Credit Risk Assessment on a Private Label Credit Card Application
  • 5.4. 2011 Kaggle competition Give Me Some Credit