recommendation-system

Simple Recommendation System in Python3+ (Using Collaborative Filtering)


Keywords
Recommendation, System, Movie, recommender, collaborative, filtering, low, rank, matrix, factorization
License
MIT
Install
pip install recommendation-system==1.0.0

Documentation

Simple Recommendation System in Python3+ (Using Collaborative Filtering)

Installation

  • pip install recommendation_system

Example Usage

from recommendation_system import complete_recommendation_table
import pandas as pd
Y_df = pd.DataFrame({'Bob': [5, '?', 4], 'Cathy': [5, 4, '?'], \
                    'Dave' : [2, 5, 5]}, index=['Toy Story',  \
                    'Despicble Me', 'Spiderman'])
output = complete_recommendation_table(Y_df, len(Y_df) + 1, \
            unknown='?', max_value=5, min_value=0, regularization_coeff=0.2)
print(output)

Example Output

  • Assume an input table that looks like this:
  • |Movie /User   | Bob | Cathy | Dave |
  • | Toy Story        |  5    |  5         |  2       |
  • | Despicble Me |  ?    |  4         |  5       |
  • | Spiderman      |  4   |  ?          |  5       |
  • The output after filling in the '?' would look something like this:
  • |Movie /User   | Bob                | Cathy                | Dave               |
  • | Toy Story        |  4.808920    |  4.917348         |  2.118293       |
  • | Despicble Me |  3.998761    |  3.874179         |  4.816824       |
  • | Spiderman      |  3.818917   |  3.744278          |  4.857076       |
  • The function 'complete_recommendation_table' fills all unknown values('?') with the predictions