Significance tests or confidence intervals: which are preferable for the comparison of classifiers?
Daniel P. BerrarJosé Antonio LozanoPublished in: J. Exp. Theor. Artif. Intell. (2013)
Keyphrases
- confidence intervals
- significance tests
- test set
- roc curve
- markov chain
- sample size
- training data
- support vector
- training set
- stochastic systems
- stopping rules
- hypothesis testing
- decision trees
- limit theorems
- hypothesis tests
- monte carlo
- training examples
- statistical analysis
- error rate
- probabilistic model
- active learning
- reinforcement learning
- search engine