Skip to content

Latest commit

 

History

History
27 lines (22 loc) · 986 Bytes

README.md

File metadata and controls

27 lines (22 loc) · 986 Bytes
Raw

Sentiment analysis on nltk movie reviews data set achieving more than 93% accuracy

Introduction

The main idea behind this project was to determine the sentiment (positive or negative) expressed by a movie review. In this project, I explored different types of feature models like the Top-N feature model and the bag-of-words model and brought them together.

Requirements

  • Python (> 3.0)
  • Pillow (comes with python by default)
  • NLTK

Description

Training files along with pretrained models of the following models are added:

  • Top-N feature model (accuracy: 0.807)
  • Unigram model (accuracy: 0.703)
  • Bigram model (accuracy: 0.786)
  • Ngram model (accuracy: 0.912)

Results of all the models are added.

References

  1. http://blog.chapagain.com.np/python-nltk-sentiment-analysis-on-movie-reviews-natural-language-processing-nlp/
  2. https://ataspinar.wordpress.com/2016/02/01/sentiment-analysis-with-bag-of-words-part-2/
  3. https://www.nltk.org/book/