--- Large Data Analysis Course Project ---
This folder is a set of simplified python codes which use sklearn package to classify movie reviews.
Two classifiers were used: Naive Bayes and SVM. SVM gives an accuracy of about 87.5%, which is slightly higher than 86% given by Naive Bayes.
imdbReviews.py
generates *.pkl
files which are the training and testing datasets.
First, set the dataset directory in the imdbReviews.py
, then run the code.
python imdbReviews.py
You will get two *.pkl
files which are needed for naive.py
and svm.py
.
To do prediction, run the following command.
python naive.py
last modified 04/16/2016