JMARS

Jointly Modelling Aspects, Ratings and Sentiments for Movie Recommendation (JMARS)

To run the code, download the data file from the following link:

and store in a folder named 'data'

Then run jmars.py as a python file:

python jmars.py

File Descriptions:

constants.py - Contains constants and global variables

indexer.py - Contains code to read imdb data and extract relevant information

optimize.py - Contains code to run optimization needed in the M-Step

sampler.py - Contains code to run Gibbs Sampling needed in the E-Step

jmars.py - Contains main code which uses the other modules and runs Gibbs Expectation-Maximization to output predicted ratings