User guide - Maintaining Personalized Experiences with Machine Learning

User guide

To start using the Maintaining Personalized Experiences with Machine Learning on AWS solution, first determine the name of the Amazon S3 bucket deployed by the AWS solution. Record the PersonalizeBucketName value under the outputs tab of the CloudFormation stack deployed in Step 1. Launch the stack. You can use the AWS Management Console to copy configuration –0—json files to your PersonalizeBucket bucket.

By default, the AWS solution detects configuration files uploaded to subfolders under the train folder in the PersonalizeBucket deployed by the solution. Configuration files must be valid JSON and their file names must end in –0—json.

Note

When managing multiple dataset groups, you must use different subfolders for each dataset group’s configuration. For example, for maintaining two dataset groups named customer_1 and customer_2, you can create configuration files as follows:

  • s3://<personalize_bucket_name>/train/customer_1/customer_1_config.json

  • s3://<personalize_bucket_name>/train/customer_2/customer_2_config.json

This allows you to keep the input datasets for each dataset group separate.