Share via

Using personalizer for various types of recommendations

deag_ranak 26 Reputation points
2021-02-11T14:13:41.353+00:00

Hello,

Azure novice here trying to set up one of your cognitive services for a POC run in a company.

I was wondering what is the recommendation scope of one personalizer. Let's suppose that there are several types of recommendations (user-user, or user-item) that I am interested in. Both of these types can have completely different rankableActions and ways of coming up with a reward number. Would it make sense to make a 'factory of personalizers', or would you recommend to entrust different types of recommendations to same instance?

Azure AI Personalizer
Azure AI Personalizer
An Azure artificial intelligence service that enables applications to personalize user experiences by learning from collective real-time user behavior.
34 questions
Azure AI services
Azure AI services
A group of Azure services, SDKs, and APIs designed to make apps more intelligent, engaging, and discoverable.
3,004 questions
Accepted answer
  1. Ramr-msft 17,741 Reputation points
    2021-02-13T01:39:38.023+00:00

    @deagranak-8569Thanks, We would recommend Use one personalization resource/ loop for every place in the ui where you want to influence decisions.
    In an e-commerce site, you may end up with, for example:

    • a loop for the home page that takes the output of a reco engine and highlights a product
    • a loop for highlighting an item in your basket to drive people to “finish the checkout”
    • a loop for “you may also want”
      Think of each personalizer loop as an AI driven sort that learns to drive a specific user decision. - even if the product information and user meta-data is the same across two loops, the models may end up being very diff as what drives rewards for one scenario will not be the same for another
      The product types and metadata and user metadata may vary for products going through the same loop.
      Remember that Personalizer is a last step ranker; that uses meta-data about actions (eg product) and users.
    1 person found this answer helpful.

0 additional answers

Sort by: Most helpful
Most helpful Newest Oldest

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.