Runs 10 iterations of 10-fold crossvalidation on the supplied arff file(s) using weka.classifiers.trees.RandomForest, and FastRandomForest, and prints accuracy, AUC (averaged over all classes) and computation time to stdout.
Stores a dataset that in FastRandomTrees use for training.
Based on the "weka.classifiers.trees.RandomForest" class, revision 1.12, by Richard Kirkby, with minor modifications: uses FastRfBagger with FastRandomTree, instead of Bagger with RandomTree. stores dataset header (instead of every Tree storing its own header) checks if only ZeroR model is possible (instead of each Tree checking) added "-threads" option
Utility functions for sorting float (single-precision) arrays, and for normalizing double arrays.
Evaluate the merit of each attribute using a random forest.
Functions used for finding best splits in FastRfTree.
Used to retrieve the out-of-bag vote of an ensemble classifier for a single instance.
Used to retrieve the out-of-bag vote of an FastRandomForest classifier for a single instance from a DataCache the forest was trained on.
Copyright © 2015–2017 Fiji. All rights reserved.