Synopsis
JMySVMLearner provides an internal Java implementation of the mySVM by Stefan Rueping.
Description
This learner uses the Java implementation of the support vector machine mySVM by Stefan RĂ¼ping. This learning method can be used for both regression and classification and provides a fast algorithm and good results for many learning tasks.
Input
- training set: expects: ExampleSet
Output
- model:
- estimated performance:
- weights:
- exampleSet:
Parameters
- kernel type: The SVM kernel type
- kernel gamma: The SVM kernel parameter gamma.
- kernel sigma1: The SVM kernel parameter sigma1.
- kernel sigma2: The SVM kernel parameter sigma2.
- kernel sigma3: The SVM kernel parameter sigma3.
- kernel shift: The SVM kernel parameter shift.
- kernel degree: The SVM kernel parameter degree.
- kernel a: The SVM kernel parameter a.
- kernel b: The SVM kernel parameter b.
- kernel cache: Size of the cache for kernel evaluations im MB
- C: The SVM complexity constant. Use -1 for different C values for positive and negative.
- convergence epsilon: Precision on the KKT conditions
- max iterations: Stop after this many iterations
- scale: Scale the example values and store the scaling parameters for test set.
- calculate weights: Indicates if attribute weights should be returned.
- return optimization performance: Indicates if final optimization fitness should be returned as performance.
- L pos: A factor for the SVM complexity constant for positive examples
- L neg: A factor for the SVM complexity constant for negative examples
- epsilon: Insensitivity constant. No loss if prediction lies this close to true value
- epsilon plus: Epsilon for positive deviation only
- epsilon minus: Epsilon for negative deviation only
- balance cost: Adapts Cpos and Cneg to the relative size of the classes
- quadratic loss pos: Use quadratic loss for positive deviation
- quadratic loss neg: Use quadratic loss for negative deviation
- estimate performance: Indicates if this learner should also return a performance estimation.