Chris Pollett > Students > Padmashali

    Print View



    [CS 297 Proposal]

    [Deliverable 1]

    [Netflix Recommender System - PDF File]

    [Deliverable 2]

    [Deliverable 3]

    [Deliverable 4]

    [CS 297 Report - PDF File]

    [CS 298 Proposal]

    [CS 298 Report - PDF File]

    [CS 298 Final Presentation Slides - PDF File]

Yelp Recommender System.


The Yelp Recommender System predicts user ratings for a particular business/restaurant based on user reviews. The dataset was provided by Yelp. You can download the dataset from the following website:
Yelp Dataset

I have made use of only the Yelp reviews data to predict ratings. The yelp_academic_dataset_review.json file has 2.7M reviews and 649K tips by 687K users for 86K businesses.

I have made use of the following features - "business_id", "user_id", "stars" for prediction. I have tried to predict by computing the business/restaurant and user biases by aggregating the data. I have calculated the biases of users and businesses with respect to the overall average rating. This recommender system tries to leverage the fact that some users have a habit of rating poorly, while some rate generously. Similarly, with the restaurants, some restaurants have a higher restaurant quality than other reataurants.

Example: Suppose we want to predict what user "Sarika" would rate the restaurant "In-n-Out Burger" and the overall average rating = 3.7, the users and the restaurants bias comes out to be 1.7 and -1.2 respectively then the predicted rating would be 3.7 + 1.7 -1.2 = 4.2.
  1. Yelp Recommender System. Implementation for the Yelp Recommender system in Python. It predicts the ratings and also generates two cvs files - user_bias.csv and restaurant_bias.csv where the users and restaurants bias can be found.
  2. Yelp Dataset can be downloaded from the website. We only require the yelp_academic_dataset_review.json file.
  • Running the code One can run the program from the terminal by running the following command:

    $ python "User_ID" "Business/Restaurant ID"

    Example:$ python 7KoVg5QMjYu8taLFSE7hNA mYSpR_SLPgUVymYOvTQd_Q
  • You can go through the following presentation to understand the math behind the working of this algorithm: Netflix Recommender System Presentation