04 | Content-Based Filtering & Hybrid: MatchBox Recommender in AzureML

Sign in to queue

Description

In the second but last module, I look into the other popular recommendation approach: content-based filtering, extending it to a hybrid recommender later on. Here, I use the MatchBox recommender provided in AzureML.

[00:48] Content-based filtering

[04:10] Demo: MatchBox recommender - training the content-based filtering model

[11:56] Demo: MatchBox recommender - deploying the content-based filtering model as a web service

[16:07] Demo: MatchBox recommender - extend to hybrid approach (training)

[27:27] Demo: MatchBox recommender - deploy the hybrid recommender as a web service

[31:45] Wrapping up.

The accompanying step-by-step guide can be found in the GitHub repository Recommenders in Azure.

This Channel9 video is part of a free online course on Microsoft Virtual Academy: Building Recommendation Systems in Microsoft Azure.

Embed

Download

Download this episode

The Discussion

  • User profile image
    Marc

    Why does my train matchbox recommender gives an error 0034 : More than one rating for user "userid" and item "Itemid" in rating prediction data table. While i don't have any duplicates in my dataset ?

  • User profile image
    Marc

    Nevermind don't know why but my split data generated duplicate data so removed them with the duplicate remover module

Add Your 2 Cents