NAME
Algorithm::LibLinear::Model
SYNOPSIS
use Algorithm::LibLinear;
my $data_set = Algorithm::LibLinear::DataSet->load(fh => \*DATA);
my $classifier = Algorithm::LibLinear->new->train(data_set => $data_set);
my $classifier = Algorithm::LibLinear::Model->load(filename => 'trained.model');
my @labels = $classifier->class_labels;
if ($classifier->is_oneclass_model) { ... }
if ($classifier->is_probability_model) { ... }
if ($classifier->is_regression_model) { ... }
say $classifier->num_classes; # == @labels
say $classifier->num_features; # == $data_set->size
for my $label (1 .. $classifier->num_classes) {
print 'Coeffs: ';
print join(' ', map {
$classifier->coefficient($_, $label);
} 1 .. $classifier->num_features);
print "\t";
print 'Bias: ', $classifier->bias($label);
print "\n";
}
my $class_label = $classifier->predict(feature => +{ 1 => 1, 2 => 1, ... });
my @probabilities = $classifier->predict_probability(feature => +{ 1 => 1, 2 => 1, ... });
my @values = $classifier->predict_values(feature => +{ 1 => 1, 2 => 1, ... });
$classifier->save(filenmae => 'trained.model');
__DATA__
+1 1:0.708333 2:1 3:1 4:-0.320755 5:-0.105023 6:-1 7:1 8:-0.419847 9:-1 10:-0.225806 12:1 13:-1
-1 1:0.583333 2:-1 3:0.333333 4:-0.603774 5:1 6:-1 7:1 8:0.358779 9:-1 10:-0.483871 12:-1 13:1
+1 1:0.166667 2:1 3:-0.333333 4:-0.433962 5:-0.383562 6:-1 7:-1 8:0.0687023 9:-1 10:-0.903226 11:-1 12:-1 13:1
...DESCRIPTION
This class represents a classifier or an estimated function generated as a return value of Algorithm::LibLinear's train method.
If you have model files generated by LIBLINEAR's train command or this class's save method, you can load them.
METHOD
Note that the constructor new is not a part of public API. You can get a instance via Algorithm::LibLinaear->train. i.e., Algorithm::LibLinear is a factory class.
load(filename => $path)
Class method. Loads a LIBLINEAR's model file and returns an instance of this class.
bias([$index])
Returns value of the bias term corresponding to the $index-th class. In case of one-class SVM (i.e., when is_oneclass_model is true,) the $index is ignored.
Recall that a trained model can be represented as a function f(x) = W^t x + b, where W is a F x C matrix, b is a C-sized vector and C and F are the numbers of classes and features, respectively. This method returns b($index) in this notation.
Note that $index is 1-based, unlike LIBLINEAR's get_decfun_bias() function.
class_labels
Returns an ArrayRef of class labels, each of them could be returned by predict and predict_values.
coefficient($feature_index, $label_index)
Returns value of the coefficient of classifier matrix. i.e., W($feature_index, $label_index) (see bias method description above.)
Be careful that both indices are 1-based just same as bias.
is_oneclass_model
Returns true if the model is trained for one-class SVM, false otherwise.
is_probability_model
Returns true if the model is trained for logistic regression, false otherwise.
is_regression_model
Returns true if the model is trained for support vector regression (SVR), false otherwise.
num_classes
The number of class labels.
num_features
The number of features contained in training set.
predict(feature => $hashref)
In case of classification, returns predicted class label.
In case of regression, returns value of estimated function given feature.
predict_probabilities(feature => $hashref)
Returns an ArrayRef of probabilities of the feature belonging to corresponding class.
This method will raise an error if the model is not a classifier based on logistic regression (i.e., not $classifier->is_probability_model.)
predict_values(feature => $hashref)
Returns an ArrayRef of decision values of each class (higher is better).
save(filename => $path)
Writes the model out as a LIBLINEAR model file.