org.apache.mahout.classifier.discriminative
Class PerceptronTrainer
java.lang.Object
org.apache.mahout.classifier.discriminative.LinearTrainer
org.apache.mahout.classifier.discriminative.PerceptronTrainer
public class PerceptronTrainer
- extends LinearTrainer
Implements training according to the perceptron update rule.
Constructor Summary |
PerceptronTrainer(int dimension,
double threshold,
double learningRate,
double init,
double initBias)
|
Method Summary |
protected void |
update(double label,
Vector dataPoint,
LinearModel model)
Implement this method to match your training strategy. |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
PerceptronTrainer
public PerceptronTrainer(int dimension,
double threshold,
double learningRate,
double init,
double initBias)
update
protected void update(double label,
Vector dataPoint,
LinearModel model)
- Implement this method to match your training strategy. Perceptron update works such that in case the predicted label
does not match the real label, the weight vector is updated as follows: In
case the prediction was positive but should have been negative, the weight vector
is set to the sum of weight vector and example (multiplied by the learning rate).
In case the prediction was negative but should have been positive, the example
vector (multiplied by the learning rate) is subtracted from the weight vector.
- Specified by:
update
in class LinearTrainer
- Parameters:
label
- the target label of the wrongly classified data point.dataPoint
- the data point that was classified incorrectly.model
- the model to update.
Copyright © 2008-2010 The Apache Software Foundation. All Rights Reserved.