public class LossMultiClassDecoder extends Object implements MultiClassDecoder, Serializable
Loss-based Multi-class Decoder
A multi-class decoder which predict the category of an example by minimizing the loss. More details can be found in the
following paper:
Allwein, E.L., Schapire, R.E., and Singer, Y., "Reducing multiclass to binary: A unifying approach for margin classifiers,"
Journal of Machine Learning Research, 1:113:C141, 2000.
Copyright: Copyright (c) 2005
Company: IST, Drexel University
| Constructor and Description |
|---|
LossMultiClassDecoder(LossFunction lossFunc) |
| Modifier and Type | Method and Description |
|---|---|
int |
decode(CodeMatrix matrix,
double[] binClassifierResults)
Predicts the label of the example given the code matrix and the results of all binary classifiers
|
int[] |
rank()
Rank all class labels
|
public LossMultiClassDecoder(LossFunction lossFunc)
public int decode(CodeMatrix matrix, double[] binClassifierResults)
MultiClassDecoderdecode in interface MultiClassDecodermatrix - the code matrix which describes the relationship between binary classifiers and categoriesbinClassifierResults - the predictions of all binary classifierspublic int[] rank()
MultiClassDecoderrank in interface MultiClassDecoderCopyright © 2018 JULIE Lab, Germany. All rights reserved.