Comments on "Leave-Cluster-Out Cross-Validation Is Appropriate for Scoring Functions Derived from Diverse Protein Data Sets": Significance for the Validation of Scoring Functions.
Pedro J. BallesterJohn B. O. MitchellPublished in: J. Chem. Inf. Model. (2011)
Keyphrases
- scoring functions
- cross validation
- data sets
- model selection
- scoring function
- training set
- unseen data
- hyperparameters
- support vector
- error estimates
- permutation tests
- generalization error
- similarity measure
- cross validated
- information criterion
- clustering algorithm
- nearest neighbor classifiers
- parameter estimation
- maximum likelihood
- semi supervised
- feature selection
- machine learning