Optimizing debt collections using constrained reinforcement learning.
Naoki AbePrem MelvilleCezar PendusChandan K. ReddyDavid L. JensenVince P. ThomasJames J. BennettGary F. AndersonBrent R. CooleyMelissa KowalczykMark DomickTimothy GardinierPublished in: KDD (2010)
Keyphrases
- reinforcement learning
- function approximation
- information retrieval
- digital libraries
- data sets
- reinforcement learning algorithms
- document collections
- learning algorithm
- optimal policy
- short term
- markov decision processes
- data mining
- model free
- metadata
- learning process
- long term
- state space
- machine learning
- robotic control
- optimal control
- action selection
- constrained problems
- heterogeneous collections
- robot control
- historical data
- learning problems
- neural network