Choosing the item interaction data used for training

To filter records used in a training dataset by event value or event type

1. Create an Interaction schema and include the EVENT_TYPE and EVENT_VALUE fields using "name" and "type" key-value pairs as shown.

   import boto3
   import json
    
    
   personalize = boto3.client('personalize')
    
   # Create a name for your schema
   schema_name = 'YourSchemaName'
    
   # Define the schema for your dataset
   schema = {
       "type": "record",
       "name": "Interactions",
       "namespace": "com.amazonaws.personalize.schema",
       "fields": [
           {
               "name": "USER_ID",
               "type": "string"
           },
           {
               "name": "ITEM_ID",
               "type": "string"
           },
           {
               "name": "EVENT_VALUE",
               "type": "float"
           },
           {
               "name": "EVENT_TYPE",
               "type": "string"
           },
           { 
               "name": "TIMESTAMP",
               "type": "long"
           }
       ],
       "version": "1.0"
   }
    
   # Create the schema for Amazon Personalize
   create_schema_response = personalize.create_schema(
       name = schema_name,
       schema = json.dumps(schema)
   )
    
   #To get the schema ARN, use the following lines
   schema_arn = create_schema_response['schemaArn']
   print('Schema ARN:' + schema_arn )

2. Format your input data to match your schema. For a code sample, see Data format guidelines.

3. Upload your data to an Amazon Simple Storage Service (Amazon S3) bucket. For a code sample, see Uploading to an Amazon S3 bucket.

4. Import your data into Amazon Personalize with the CreateDatasetImportJob API. Be sure to record your dataset group Amazon Resource Name (ARN) because you will need it when you create the solution. For a code sample, see Importing bulk records (Amazon SDKs).

5. Get the ARN of the recipe that you want to use when you create your solution. You'll need it when you create the solution.

   # Display the ARNs of the recipes
   recipe_list = personalize.list_recipes()
   for recipe in recipe_list['recipes']:
       print(recipe['recipeArn'])
       
   # Store the ARN of the recipe that you want to use
    recipe_arn = "arn:aws:personalize:::recipe/aws-recipe-name"

6. Call the CreateSolution API. If you want to specify the event type, for example “purchase”, set it in the eventType parameter. If you want to specify an event value, for example 10, set it in the eventValueThreshold parameter. You can also specify both an event type and an event value.

   # Create the solution
   create_solution_response = personalize.create_solution(
       name = "your-solution-name",
       datasetGroupArn = dataset_group_arn,
       recipeArn = recipe_arn,
       eventType = 'watched',
       solutionConfig = {
           "eventValueThreshold": "0.5"
       }
   )
   
   # Store the solution ARN
   solution_arn = create_solution_response['solutionArn']
   
   # Use the solution ARN to get the solution status
   solution_description = personalize.describe_solution(solutionArn = solution_arn)['solution']
   print('Solution status: ' + solution_description['status'])

7. When you have the solution, use it to train a model by specifying its solution ARN in a CreateSolutionVersion request.

   # Create a solution version
   create_solution_version_response = personalize.create_solution_version(solutionArn = solution_arn)
   
   
   # Store the solution version ARN
   solution_version_arn = create_solution_version_response['solutionVersionArn']
       
   # Use the solution version ARN to get the solution version status.
   solution_version_description = personalize.describe_solution_version(
       solutionVersionArn = solution_version_arn)['solutionVersion']
   print('Solution version status: ' + solution_version_description['status'])