Recommender-System

Implementation of various recommendation algorithms such as Collaborative filtering, SVD and CUR-decomposition to predict user movie ratings

Requirements

• python
• numpy
• pandas
• scipy

Run

The dataset used for training MovieLens data sets were collected by the GroupLens Research Project. This data set consists of 100,000 ratings ranging between 1 to 5 from 943 users on 1682 movies. You can download it from Movielens 100K link.

To run, execute following commands

python3 recommender.py

Output

CollaborativeFiltering RMSE 0.8452675603293487
CollaborativeFiltering Spearmans Rank Correlation 0.9999999240526046
CollaborativeFiltering Precision On TopK 0.9024871794603606

CollaborativeFilteringBaseline RMSE 0.7675328677210587
CollaborativeFilteringBaseline Spearmans Rank Correlation 0.9999999373792241
CollaborativeFilteringBaseline Precision On TopK 0.8839798773318012

SingularValueDecomposition RMSE 3.780525477770839e-13
SingularValueDecomposition Spearmans Rank Correlation 1.0
SingularValueDecomposition Precision On TopK 1.0

CUR RMSE 1.7448751755958805
CUR Spearmans Rank Correlation 0.999999676366695
CUR Precision On TopK 1.0

Various recommendation techniques formulae

Collaborative Filtering

Collaborative Filtering using Baseline Estimate

Singular Value Decomposition

CUR Decomposition

Error Calculations

License

Distributed under the Apache-2.0 License. See LICENSE.txt for more information.