Offline Evaluation for Reinforcement Learning-based Recommendation: A Critical Issue and Some Alternatives.
Romain DeffayetThibaut ThonetJean-Michel RendersMaarten de RijkePublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- decision makers
- function approximation
- real time
- data mining
- evaluation model
- case study
- recommendation systems
- evaluation method
- optimal policy
- reinforcement learning algorithms
- evaluation methods
- gold standard
- evaluation criteria
- evaluation metrics
- dynamic programming
- recommender systems
- e learning
- neural network