Administer > Manage WebSphere Commerce features > Marketing > Collaborative filtering > WebSphere Commerce Recommendation Engine powered by LikeMinds
Background LikeMinds utilities
LikeMinds depends on two utilities to update the database:
- sifter
- This utility runs continuously to identify mentors for new users and recomputes the best set of mentors for existing users. The mentor set for a user may change as the sifter gathers more information about the user or the mentors.
- buildstats
- This utility runs once a day to update statistics for each item, such as the number of ratings or transactions, the average rating, the standard deviation in the average rating, and the default recommendation information.
These utilities are run from the LikeMinds utility servlet.
You must configure these utilities to run at appropriate times.
- Sifter utility
- buildstats utility
The buildstats utility updates database statistics when you use LikeMinds.
- Configure when the buildstats utility runs
- Start the LikeMinds server
- Stop the LikeMinds server
- Create the LikeMinds database schema
LikeMinds Personalization Server collaborative filtering technology complements the existing WebSphere Commerce personalization offering. It consists of a client component, a WebSphere Application Server application, and a database.
- Configure the LikeMinds Personalization Server
You must configure the LikeMinds Personalization Server in WebSphere Application Server.
- Configure WebSphere Commerce to interact with LikeMinds
To enable WebSphere Commerce Server to access the LikeMinds database configure WebSphere Commerce in WebSphere Application Server. LikeMinds cannot be installed in WebSphere Commerce Developer.
- Enable the LikeMinds listener in WebSphere Commerce
The final step is to enable the LikeMinds listener within WebSphere Commerce.
- Test LikeMinds Personalization Server
- Specify the minimum number of Clickstream activities for a user
- How the sifter selects mentors
- Verify that the sifter is functioning
Related concepts