Machine-learning scoring functions trained on complexes dissimilar to the test set already outperform classical counterparts on a blind benchmark.
Hongjian LiGang LuKam-Heung SzeXianwei SuWai-Yee ChanKwong-Sak LeungPublished in: Briefings Bioinform. (2021)
Keyphrases
- test set
- scoring functions
- training set
- machine learning
- scoring function
- error rate
- decision trees
- active learning
- supervised learning
- test data
- training data
- evaluation methodology
- machine learning algorithms
- data sets
- computer vision
- learning algorithm
- feature selection
- text mining
- syntactic dependencies
- support vector machine
- classification accuracy
- support vector
- similarity measure