SVM Apply Function

This function applies a Support Vector Machine (SVM) model previously trained by a SVM procedure to a feature vector, predicting a label or returning a regression result. It is a wrapper around the popular open-source LIBSVM library. For more information about LIBSVM: http://www.csie.ntu.edu.tw/~cjlin/libsvm/

Configuration

A new function of type svm named <id> can be created as follows:

mldb.put("/v1/functions/"+<id>, {
    "type": "svm",
    "params": {
        "modelFileUrl": <Url>
    }
})

with the following key-value definitions for params:

Input and Output Values

The SVM function of this type has a single input value called embedding which is a feature vector corresponding to the feature vector that was used for training the SVM model, without the label.

The function outputs a single value output which depend on the type of SVM trained:

• classification and nu-classification will return the predicted label of the feature vector
• one class will return a value corresponding to how alike the feature vector is to the training data
• regression and nu-regression will return the regression value

See also

• The classifier.test procedure type allows the accuracy of a predictor to be tested against held-out data.
• The svm.train procedure type trains an SVM.