The offline metrics provided by Personalize are meant to provide an overall directional sense of how the solution version performed against held out data from your interactions dataset. If you want to do your own inspection of recommendations generated by a solution version, you can create a batch inference job across a sample of users (or all users). Then you can inspect and analyze the recommendations in the output file across the users in the input file.
If you're looking to identify users that have a affinity for an item or item attributes, take a look at the new user-segmentation feature.
What is the recommended way of handling player authentication using Cognito?Accepted Answerasked 3 years ago
Is there a way to pull a list of users who have access to each folder and file under a parent folder?asked 5 months ago
Looking for an easy way to bulk delete SQS queues and SNS topicsasked 6 months ago
Is there any way to download the recommended lists of users as bulk when using AWS-Personalize (specifically using user-personalization)asked a year ago
Is there a way to improve AWS Workdocs download speeds ?asked 6 months ago
Multi Download from Amazon S3 (Glacier Deep Archive)asked a month ago
Personalize Recommendations for Anonymous Users using SessionIdasked 2 months ago
Is there a way to logically group steps within a recipe?
How to Download any file from s3 using the pyspark kernel in emr notebook ?asked 2 years ago
Is there a way to add a step between other steps in a recipe with the UI?