fasttext

Text/Sentence Classification Python Module

Installation:

pip install fasttext

Training Data Set Format:

Example: __label__moviereview The film's two major strengths come down to the two most important ingredients - cast and story.

Explanation:

Prefix: label

Label_name: moviereview

Sentence: The film's two major strengths come down to the two most important ingredients - cast and story.

Note: There should be space between Label_name and Sentence and there should be one Sentence/Line followed by it's Label_name.

Training the Data Set:

classifier = fasttext.supervised('train_data.txt', 'model_name', label_prefix ='__label__')

Testing the Data Set:

test_result = classifier.test('test_data.txt')

print('Precision', test_result.precision)

print('Recall', test_result.recall)

print('No. of examples:', test_result.nexamples)

Loading the model:

model = fasttext.load_model('model_directory_path',label_prefix='__label__')

Prediction:

Sentence = "provide any sentence for testing"

predict = (model.predict(sentence))

Prediction with probability:

predict_prob = model.predict_proba(sentence, k=2)

Note: k value indicates no. of label predictions.