Skip to content

Latest commit

 

History

History
28 lines (20 loc) · 874 Bytes

README.md

File metadata and controls

28 lines (20 loc) · 874 Bytes

WebApp recommend article_ids

This app can recommend the top 5 article_ids from Globo.com user_ids.

Context

This was a kaggle competition held several years ago. See News Portal User Interactions for much more detailed documentation.

About the app

It is in fact a python BayesianPersonalizedRanking model using an Azure serverless function.

Librairies

You need to install first some python librairies:

Requests
streamlit

Goal

As you know now, our main objective is to recommend for users the 5 articles that they may be interested to read.

Target

  • Top 5 article_ids

Our function recommand_article gives us a list of the top 5 article_ids.

Recommandations

I strongly recommand you to install packages from the requirements.txt.