AI Simplified: SHAP Values in Machine Learning
What are SHAP values and how are they being used to explain machine learning predictions?
Meet Mark Romanowsky, a Data Scientist at DataRobot. Mark explains the important role that SHAP values play in machine learning by providing real-world examples: how a group of friends can fairly and efficiently share an Uber ride to their separate homes, and helping college students at risk of not graduating.
In everyday life, Shapley values are a way to fairly split a cost or payout among a group of participants who may not have equal influence on the outcome. In machine learning models, SHAP values are a way to fairly assign impact to features that may not have equal influence on the predictions.
Learn more in his AI Simplified video:
Ready for more AI Simplified? Check out these other videos:
- AI Simplified: Model Evaluation
- AI Simplified: Trusting Your AI
- AI Simplified: Model Interpretability
-
Automating Model Risk Compliance: Model Validation
May 26, 2022· 9 min read -
AI for Climate Change and Weather Risk
May 25, 2022· 3 min read -
AI Applications for Border Transportation
May 23, 2022· 3 min read
Latest posts