IMDb Sentiment Analysis

Machine learning methods to predict the positivity/sentiment of an IMDb movie review.

Setup

1. Clone repository (run git clone https://github.com/RylanSteinkey/imdbSentimentAnalysis.git)
2. Change directories into the project folder: cd imdbSentimentAnalysis
3. Download anaconda or miniconda (python 3.7), instructions for that are here
4. Install dependecies: run conda env create -f envi.yaml
5. Run snakemake
6. Check results.txt for accuracy and a ranked list of important words

Other Machine Learning Models

After the above has successfully run, you can execute other models by running models.py as:

python models.py XGB -- XGBoost (68.2% accuracy with 1000 samples)

python models.py SVM -- Support Vector Machine (68.4% accuracy with 1000 samples)

python models.py MNB -- Multinomial Naive Bayes (68.4% accuracy with 1000 samples)

python models.py ANN -- Artificial Neural Network (78.8% accuracy with 1000 samples)