Movie Recommender System Based on Paragraph to Vector

One Paragraph of project description goes here

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

Python (version from 2.7 to 3.6 will do) and PostgreSQL (or any powerful and efficient open source database).

Built With

• gensim - The web framework used
• nltk - Dependency Management

Data Manipulation

Data Collection

Download the data source and get account for API (see Data Source part). Dump the data into database and start data collection from Amazon.

Data Cleansing

Recorded in the sql and Python scripts. See blog for details.

Model Training

Both single worker version (which eliminate randomness) and the fast version are included. See blog for details.

Running the tests

The are various way of testing the model, but none of which are included.

Results

Results are dumped into file by the script.

Deployment

Deployed in AWS with a whole website.

Versioning

First version of the model.

Authors

• Eric Kehan Pan - Initial work - PurpleBooth

See also the list of contributors who participated in this project.

Data Source

• Amazon movie reviews - 8 million movie reviews
• Amazon Advertising API - Source for movie names and posters

Acknowledgments

• J. McAuley and J. Leskovec. From amateurs to connoisseurs: modeling the evolution of user expertise through online reviews. WWW, 2013.