javifmz/lexicon-based-classifier

Classifies texts using a previously generated lexicon.

Lexicon-based Classifier

Classifies texts using a previously generated lexicon. The lexicon contains a list of weighted terms for each of the classification labels. The weights of these terms are added to each label, and the label with the higher weight is chosen.

Lexicon format example

{
  "positive": [
    { "term": "excellent", "weight": 1.0 },
    { "term": "good", "weight": 0.7 }
  ],
  "negative": [
    { "term": "horrible", "weight": 1.0 },
    { "term": "bad", "weight": 0.8 }
  ],
}

Basic usage

classifier = LexiconBasedClassifier(lexicon)
prediction = classifier.predict(X)

Usage with different prediction indexes

labels = [ "positive", "negative" ]
classifier = LexiconBasedClassifier(lexicon, labels=labels)
prediction = classifier.predict(X)

Execute the example

python -m lbclassifier_example