aakashrkaku / Yelp-recommender-system

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Yelp-Recommendation-System

Download Yelp's Business.CSV and Review.CSV to run this code on your machine. Code along with the output can be viewed in .ipynb file

Winner of “Best First Year Project” award at NYU CDS Academy Awards.

  • Built a recommendation engine for recommending restaurants to Yelp users using traditional models like Cosine similarity based model, SVD and Alternating Least Square model; Rating Matrix was very sparse with sparsity of 99.4%
  • Developed advanced models like Stochastic Gradient Descent - based model, Neural Network based model, Random Forest Regressor based model and an ensemble model to achieve higher performance on the sparse rating matrix completion task

Sparsity of Dataset is high

About


Languages

Language:Jupyter Notebook 100.0%