Feature space reduction method for ultrahigh-dimensional, multiclass data: random forest-based multiround screening (RFMS).
Gergely HanczárMarcell StippingerDávid HanákMarcell T. KurbuczOlivér M. TörteliÁgnes ChripkóZoltán SomogyváriPublished in: Mach. Learn. Sci. Technol. (2023)
Keyphrases
- multi class
- random forest
- data sets
- feature space
- reduction method
- input space
- feature selection
- data structure
- data analysis
- data points
- neural network
- data distribution
- training data
- cost sensitive
- decision trees
- machine learning
- multiple classes
- classification accuracy
- pairwise
- support vector machine
- bayesian networks
- test data
- random forests
- binary classification
- ensemble classifier