Recommender System

CS F469 IR Assignment - 3

Problem Statement:

The task is to compare various techniques used in implementing Recommender Systems on the basis of their errors using Root Mean Square Error, Precision on top K and Spearman Rank Correlation. Also compare their overall running time and prediction time. The techniques implemented and compared are:

1. Collaborative Filtering.
2. Collaborative Filtering using Baseline approach.
3. Singular Value Decomposition(SVD).
4. SVD with 90% energy.
5. CUR.
6. CUR with 90% energy.

About the project

Dataset used - Movie lens Dataset is used in this assignment consisting of 6040 users rating of 3883 movies

Have a look at the file Design Document. It includes the concepts used along with the time taken for each implementation step.

Project By:

• Kriti Jethlia: Email- f20180223@hyderabad.bits-pilani.ac.in
• Jui Pradhan: Email- f20180984@hyderabad.bits-pilani.ac.in
• Anusha Agarwal: Email- f20180032@hyderabad.bits-pilani.ac.in

---

How to run the code

1. Clone the repository : https://github.com/JuiP/RecommenderSystem.git

2. cd RecommenderSystem

3. Change the path to ratings.dat in the python code, wherever you have saved the ml-1m folder(preferably in the same folder) and run the python script for each method.

4. Run file:

          python3 collab_user.py
          python3 collab_global.py
          python3 cur.py
          python3 cur_90.py
          python3 svd.py
          python3 svd_90.py
   

5. The Root Mean Square Error(RMSE), Precision on top K, Spearman Rank Correlation and Time taken for prediction is printed at the end of each file execution.:)

---

Dependencies/modules used

• time
• math
• pandas
• Numpy
• random