question

GregorioRojas-6472 avatar image
0 Votes"
GregorioRojas-6472 asked romungi-MSFT commented

Training Personalizer

I am considering using Personalizer for project and have found limited third party metrics for this service. This one article indicates needing TENS of thousands of hits to get good results.

https://medium.com/@EnefitIT/we-tested-azure-personalizer-heres-what-you-can-expect-8c5ec074a28e

Can any one provide any other data?

Obviously, over time it will get better, but does it have to get to 10K+ to get good?

azure-machine-learningazure-personalizer
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

1 Answer

romungi-MSFT avatar image
0 Votes"
romungi-MSFT answered romungi-MSFT commented

@GregorioRojas-6472 The minimum requirements to have an effective recommendation is to have a minimum of ~1k/day content-related events. Higher rate of events do help you to provide faster and better recommendations. All the requirements are documented in the official documentation page of the service. The samples repo provides some data along with the quickstart's from the documentation to get started. The service now provides an E0 tier or apprentice mode that helps you test the service and gain confidence to move to a higher tier with production level recommendations.


· 2
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Thank you very much @romungi-MSFT ... but when it says "the service takes longer to determine the single best content item." ...

does it mean that it will take longer in that when I make a request for a recommendation, the api request itself, will take longer to respond in time.

or

does it mean that if it does not get 1K events per day, that it will take longer (as in it will take more days ) to get good recommendations from the api request?







0 Votes 0 ·

@GregorioRojas-6472 It takes more time to provide good recommendations. The amount of events should not effect the api response time.

0 Votes 0 ·