TWITTER SENTIMENT CLASSIFIER
=============================

This package implements an twitter sentiment classifier and server for the Bottle
Python web framework.  You can get the Bottle (and learn more about it) at

  http://bottlepy.org/

SETUP
=====

 1. Install the nltk library, version 3.0.0 or later with all packages.

 2. Install Bottle 0.12.8 or later.

 3. To run the server, run:

    python hellobottle.py CLASSIFIER_FOLDER CLASSIFIER_NAME

    where CLASSIFIER_FOLDER is folder that contains classifiers.
          Folder name can't contain any spaces.

          CLASSIFIER_NAME is a name of classifier:
          "maxent" for Maxent algorithm,
          "logreg" for Logistic Regression,
          "svc" for LinearSVC,
          "nv" for NaiveBayes.

 6. Point your web browser at the server at

    http://localhost:8080/

    to begin.



CLASSIFIERS
===========

 1. All classifiers were made by NLTK-Trainer – https://github.com/japerk/nltk-trainer

 2. All classifiers are based on the Twitter Sentiment Corpus data – http://www.sananalytics.com/lab/twitter-sentiment/