[
https://issues.apache.org/jira/browse/UNOMI-193?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Romain Gauthier updated UNOMI-193:
----------------------------------
Description:
Currently, it is possible to use Unomi for personalization, but it would be
great if it could go further and support more advanced use cases, based on
machine learning:
* Product / content recommendations: the best thing to do would be to support
both condition based recommendations as of today, but with a scoring / a sort
based on ML algorithms.
* Audiience lookalike: Unomi to suggest segments if some profiles have a
similar behaviour
* Predictive analytics: To guess how many profiles should fulfil this goal in
the next x days ?
We could also use ML to provide probabilistic identity matching.
Do you have other use cases in mind ?
Best,
Romain
> Machine learning for Apache Unomi
> ---------------------------------
>
> Key: UNOMI-193
> URL: https://issues.apache.org/jira/browse/UNOMI-193
> Project: Apache Unomi
> Issue Type: New Feature
> Reporter: Romain Gauthier
> Priority: Major
>
> Currently, it is possible to use Unomi for personalization, but it would be
> great if it could go further and support more advanced use cases, based on
> machine learning:
> * Product / content recommendations: the best thing to do would be to
> support both condition based recommendations as of today, but with a scoring
> / a sort based on ML algorithms.
> * Audiience lookalike: Unomi to suggest segments if some profiles have a
> similar behaviour
> * Predictive analytics: To guess how many profiles should fulfil this goal
> in the next x days ?
> We could also use ML to provide probabilistic identity matching.
> Do you have other use cases in mind ?
>
> Best,
> Romain
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
