Skip to main content

User article recommender system for IBM Watson Studio

Project description

Recommender_ibmws

Recommender IBMWS package is a Python Package to recommend articles for users of the IBM Watson Studio platform.

This Recommender uses a hybrid approach of Content-Based Filtering and Collaborative Filtering to make recommendations.
The content base recommendation system developed relies on a user profile, text vectorization, similarity calculation, ranking and recommendation, as well as handling of new users to deal with the cold start problem.

How to use:

Import the Recommender

from recommender_ibmws.recommender import Recommender
rec = Recommender()

Load data

rec.load_data(inter_path='user-item-interactions.csv', content_path='articles_community.csv')

Make recomendations

rec.make_content_recs(user_id=8, m=10)

Find number of users

rec.n_users

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page