A new criterion using information gain for action selection strategy in reinforcement learning.
Kazunori IwataKazushi IkedaHideaki SakaiPublished in: IEEE Trans. Neural Networks (2004)
Keyphrases
- selection strategy
- information gain
- reinforcement learning
- feature selection
- decision trees
- selection strategies
- text categorization
- chi squared
- chi square
- mutual information
- gini index
- selection algorithm
- correlation coefficient
- naive bayes
- occurrence frequency
- fitness function
- learning algorithm
- machine learning
- feature subset
- data sets
- face recognition
- document frequency
- data mining
- unsupervised learning
- pattern recognition
- genetic algorithm