GA(M)E-QSAR: A Novel, Fully Automatic Genetic-Algorithm-(Meta)-Ensembles Approach for Binary Classification in Ligand-Based Drug Design.
Yunierkis Pérez-CastilloCosmin LazarJonatan TaminauMatheus FroeyenMiguel Ángel Cabrera-PérezAnn NowéPublished in: J. Chem. Inf. Model. (2012)
Keyphrases
- drug design
- fully automatic
- binary classification
- genetic algorithm
- ensemble methods
- genetic algorithm ga
- fitness function
- semi automatic
- multi class
- support vector
- drug discovery
- learning problems
- prediction accuracy
- cost sensitive
- multi label
- protein structure prediction
- support vector machine
- benchmark datasets
- generalization error
- decision trees
- base classifiers
- machine learning methods
- evolutionary algorithm
- metaheuristic
- simulated annealing
- class imbalance
- protein protein interactions
- image classification
- feature subset
- similarity measure
- class labels
- graph cuts
- semi supervised
- machine learning