Lexicon Classifier

Source code: https://github.com/draperunner/fjlc

This package is a Python port of the Lexicon Creator and Classifier of Valerij Fredriksen and Brage Ekroll Jahren (2016). It is compatible with Python version >= 3.

The original Java code is available here: https://github.com/freva/Masteroppgave

If using this package in your publications, please cite

Valerij Fredriksen and Brage Ekroll Jahren. Twitter Sentiment Analysis: Exploring Automatic Creation of Sentiment Lexica. Master's thesis, 2016.

Installation

pip install fjlc

Lexicon Classifier

The LexiconClassifier uses the best performing lexicon of Fredriksen and Jahren. You can specify your own lexicon, see Options below.

Usage

from fjlc import LexiconClassifier
lc = LexiconClassifier()

You can classify a single tweet or a list of tweets:

lc.classify("I am happy!")  # 'POSITIVE'
lc.classify(["I am happy!", "I hate rain"])  # ['POSITIVE', 'NEGATIVE']

You can get the sentiment value of a single tweet or multiple tweets

lc.calculate_sentiment("I am happy!")  # 5.599244615570646
lc.calculate_sentiment(["I am happy!", "I hate rain"])  # [5.599244615570646, -2.767224666516315]

Options

The LexiconClassifier takes three options:

• lexicon: Path to sentiment lexicon file
• options: Path to options file
• dictionary: Path to canonical dictionary

Lexicon Creator

Usage

from fjlc import LexiconCreator
lc = LexiconCreator()

Incomplete, untested.