-
Notifications
You must be signed in to change notification settings - Fork 11
Home
LibOPF is a library of functions and programs for free usage in the design of optimum-path forest classifiers. This second version contains some additional resources related to the supervised OPF classifier reported in references [PapaIJIST09,PapaPR12], and also contains the unsupervised version of OPF reported in reference [RochaIJIST09].
A short explanation about the method can be found in http://www.ic.unicamp.br/~afalcao/libopf/. Please read the COPYRIGHT file before using LibOPF.
For large datasets (thousands/millions of samples), it is usually desirable to keep some maximum size for the training set. However, an evaluation set can improve the training samples during pseudo tests (learning procedure). Therefore, LibOPF provides a program to randomly split the dataset into training, evaluation and test sets (opf_split
).
One can project an OPF classifier by using the program opf_train
and test it by using the program opf_classify
. However, for large datasets, the program opf_learn
substitutes the opf_train
by learning from classification errors in the evaluation set without increasing the training set size. Afterwards, the classifier is tested by using opf_classify
.
In the case of time-consuming distance functions, one can generate a precomputed distance file in the format specified in section opf_distance
Usage. LibOPF also provides a program (opf_distance
) with some options of distance functions, which generates a precomputed distance file.
Current version: 3.1 (May, 2015)