NAME¶

svm-predict - make predictions based on a trained SVM model file and test data

SYNOPSIS¶

svm-predict [ -b probability_estimates ] [ -q ] test_data model_file [ output_file ]

DESCRIPTION¶

svm-predict uses a Support Vector Machine specified by a given input model_file to make predictions for each of the samples in test_data
The format of this file is identical to the training_data file used in svm_train(1) and is just a sparse vector as follows:

<label> <index1>:<value1> <index2>:<value2> . . .

.

.

.

There is one sample per line. Each sample consists of a target value (label or regression target) followed by a sparse representation of the input vector. All unmentioned coordinates are assumed to be 0. For classification, <label> is an integer indicating the class label (multi-class is supported). For regression, <label> is the target value which can be any real number. For one-class SVM, it's not used so can be any number. Except using precomputed kernels (explained in another section), <index>:<value> gives a feature (attribute) value. <index> is an integer starting from 1 and <value> is a real number. Indices must be in an ASCENDING order. If you have label data available for testing then you can enter these values in the test_data file. If they are not available you can just enter 0 and will not know real accuracy for the SVM directly, however you can still get the results of its prediction for the data point.

If output_file is given, it will be used to specify the filename to store the predicted results, one per line, in the same order as the test_data file.

OPTIONS¶

-b probability-estimates

probability_estimates is a binary value indicating whether to calculate probability estimates when training the SVC or SVR model. Values are 0 or 1 and defaults to 0 for speed.

-q

quiet mode; suppress messages to stdout.

FILES¶

training_set_file must be prepared in the following simple sparse training vector format:

<label> <index1>:<value1> <index2>:<value2> . . .

.

.

.

There is one sample per line. Each sample consist of a target value (label or regression target) followed by a sparse representation of the input vector. All unmentioned coordinates are assumed to be 0. For classification, <label> is an integer indicating the class label (multi-class is supported). For regression, <label> is the target value which can be any real number. For one-class SVM, it's not used so can be any number. Except using precomputed kernels (explained in another section), <index>:<value> gives a feature (attribute) value. <index> is an integer starting from 1 and <value> is a real number. Indices must be in an ASCENDING order.

ENVIRONMENT¶

No environment variables.

DIAGNOSTICS¶

None documented; see Vapnik et al.

BUGS¶

Please report bugs to the Debian BTS.

AUTHOR¶

Chih-Chung Chang, Chih-Jen Lin <cjlin@csie.ntu.edu.tw>, Chen-Tse Tsai <ctse.tsai@gmail.com> (packaging)

SEE ALSO¶

svm-train(1), svm-scale(1)