@PublicationReference(title="Online Passive-Aggressive Algorithms on a Budget", author={"Zhuang Wang","Slobodan Vucetic"}, year=2010, type=Conference, publication="Proceedings of the 13th International Conference on Artificial Intelligence and Statistics (AISTATS)", url="http://jmlr.csail.mit.edu/proceedings/papers/v9/wang10b/wang10b.pdf", notes="This presents the PA-I-R algorithm. This implementation is non-budgeted form. We this also implements PA-R and PA-II-R") public class OnlineRampPassiveAggressivePerceptron extends OnlinePassiveAggressivePerceptron.QuadraticSoftMargin
OnlinePassiveAggressivePerceptron.AbstractSoftMargin, OnlinePassiveAggressivePerceptron.LinearSoftMargin, OnlinePassiveAggressivePerceptron.QuadraticSoftMarginaggressiveness, DEFAULT_AGGRESSIVENESSDEFAULT_UPDATE_BIASupdateBiasvectorFactory| Constructor and Description |
|---|
OnlineRampPassiveAggressivePerceptron()
Creates a new
OnlineRampPassiveAggressivePerceptron with default parameters. |
OnlineRampPassiveAggressivePerceptron(double aggressiveness)
Creates a new
OnlineRampPassiveAggressivePerceptron with the given
aggressiveness. |
OnlineRampPassiveAggressivePerceptron(double aggressiveness,
VectorFactory<?> vectorFactory)
Creates a new
OnlineRampPassiveAggressivePerceptron with the given parameters. |
| Modifier and Type | Method and Description |
|---|---|
<InputType> |
computeUpdate(DefaultKernelBinaryCategorizer<InputType> target,
InputType input,
boolean actualCategory,
double predicted)
Compute the update weight in the linear case.
|
double |
computeUpdate(LinearBinaryCategorizer target,
Vector input,
boolean actualCategory,
double predicted)
Compute the update weight in the linear case.
|
computeUpdategetAggressiveness, setAggressivenesscomputeDecay, computeDecay, computeRescaling, computeRescaling, createInitialLearnedObject, initialize, initialize, isUpdateBias, setUpdateBias, update, updatecreateKernelLearner, learn, update, update, updatecreateInitialLearnedObject, getVectorFactory, setVectorFactory, updateupdateclone, learn, learn, updateequals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, waitupdatelearnlearncreateInitialLearnedObject, update, updateclonepublic OnlineRampPassiveAggressivePerceptron()
OnlineRampPassiveAggressivePerceptron with default parameters.public OnlineRampPassiveAggressivePerceptron(double aggressiveness)
OnlineRampPassiveAggressivePerceptron with the given
aggressiveness.aggressiveness - The aggressiveness. Must be positive.public OnlineRampPassiveAggressivePerceptron(double aggressiveness,
VectorFactory<?> vectorFactory)
OnlineRampPassiveAggressivePerceptron with the given parameters.aggressiveness - The aggressiveness. Must be positive.vectorFactory - The factory to use to create new weight vectors.public double computeUpdate(LinearBinaryCategorizer target, Vector input, boolean actualCategory, double predicted)
AbstractLinearCombinationOnlineLearnercomputeUpdate in class OnlinePassiveAggressivePerceptrontarget - Target to compute the update for.input - Input to use in computing the update.actualCategory - The actual category of the input.predicted - The predicted category of the input.public <InputType> double computeUpdate(DefaultKernelBinaryCategorizer<InputType> target, InputType input, boolean actualCategory, double predicted)
AbstractLinearCombinationOnlineLearnercomputeUpdate in class OnlinePassiveAggressivePerceptronInputType - The input value for learning.target - Target to compute the update for.input - Input to use in computing the update.actualCategory - The actual category of the input.predicted - The predicted category of the input.