| MATLAB File Help: dml.one_against_rest | View code for dml.one_against_rest | Default Topics |
one_against_rest one-against-rest binary classification.
DESCRIPTION
This class evaluates a binary classifier on all possible pairs of class
labels. Borda count is used as default combfun. Can be overridden.
EXAMPLE
X = rand(15,20); Y = [1 1 1 1 1 2 2 2 2 2 3 3 3 3 3]';
m = dml.one_against_rest('mva',dml.svm);
m = m.train(X,Y);
Z = m.test(X);
DEVELOPER
Marcel van Gerven (m.vangerven@donders.ru.nl)| Superclasses | dml.method |
| Sealed | false |
| Construct on load | false |
| one_against_rest | one-against-rest binary classification. |
| combfun | 'borda' borda count |
| indims | dimensions of the input data (excluding the trial dim and time dim in time series data) |
| mva | used multivariate analysis |
| nclasses | number of classes |
| restart | when false, starts at the previously learned parameters; needed for online learning and grid search |
| verbose | whether or not to generate diagnostic output |
| model | returns the following parameters: | |
| test | get posteriors for all pairs | |
| train |