Bandit Problems in Workflow
Jon Morra explains how bandit problems, a semi supervised learning approach, fit into various workflows. He discusses how bandit problems are used in A/B testing and other domains to determine the best actions to take based on context, action, reward, and probability. The effectiveness of bandit problems has been proven in product testing and email campaigns.In this clip
From this podcast
Data Skeptic
Data Science at eHarmony
Related Questions