Configuration

Voting Sub-classifiers

The main classifier is a Voting classifier that uses a series of sub-classifiers to classify the text, then chooses the most-popular result. The sub-classifiers used inside the Voting classifier are defined in the voting.yaml file:

voting:
    multinomilnb:
        module: sklearn.naive_bayes
        class: MultinomialNB
        description: Multinomial NB classifier from SciKit
    bernouillinb:
        module: sklearn.naive_bayes
        class: BernoulliNB
        description: Bernouilli NB classifier from SciKit
    logisticregression:
        module: sklearn.linear_model
        class: LogisticRegression
        description: Logistic Regression classifier from SciKit
    sgd:
        module: sklearn.linear_model
        class: SGDClassifier
        description: SGD classifier from SciKit
    linearrsvc:
        module: sklearn.svm
        class: LinearSVC
        description: Linear SVC classifier from SciKit

Note how there is a single voting section containing a list (actually a dictionary) of sub-classifiers. Each entry is of the form:

some label:
    module: the python module containing the classifer class
    class: the class name
    description: some textual description