Updating a recommender - Amazon Personalize
Updating a recommender (Amazon Personalize console)Updating a recommender (Amazon CLI)Updating a recommender (Amazon SDKs)
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Updating a recommender

After you create a recommender, you can update the recommender's configuration:

  • You can update the columns the recommender uses in training. If you modify the columns used when training, Amazon Personalize automatically starts a full retraining of the models backing your recommender. While the update completes, you can still get recommendations from the recommender. The recommender uses the previous configuration until the update completes. To track the status of this update, use the latestRecommenderUpdate returned in the DescribeRecommender operation. If you provide the same columns you provided when you created the recommender, no update occurs.

  • You can update the recommender's minimum recommendation requests per second. This specifies the baseline recommendation request throughput that's provisioned by Amazon Personalize. A high value will increase your bill. We recommend starting with 1. Track your usage using Amazon CloudWatch metrics, and increase it as necessary. For more information, see Minimum recommendation requests per second and auto-scaling.

  • For Top picks for you and Recommended for you use cases, you can update exploration configuration by adjusting the emphasis on exploring relevant items and the exploration item age cutoff. For information about exploration, see the section for your use case in Choosing a use case.

You can update recommenders with the Amazon Personalize console, Amazon Command Line Interface (Amazon CLI), or Amazon SDKs.

Updating a recommender (Amazon Personalize console)

After you create a recommender, you can update it. You can update the columns the recommender uses in training and the recommender's minimum recommendation requests per second. For Top picks for you and Recommended for you use cases, you can update exploration configuration. To update a recommender with the console, do the following.

To update a recommender's configuration (console)

  1. Open the Amazon Personalize console at https://console.amazonaws.cn/personalize/home and sign in to your account.

  2. On the Dataset groups page, choose your Domain dataset group.

  3. From the navigation pane, choose Recommenders.

  4. On the Recommenders page, choose the recommender that you want to update.

  5. In Recommender configuration choose Edit.

  6. Change the recommender's configuration and choose Update. For information on the different configuration options, see Creating recommenders (console).

Updating a recommender (Amazon CLI)

To update recommender with the Amazon CLI, use the update-recommender command. Provide the Amazon Resource Name (ARN) for the recommender and updated configuration. The following code shows how to update the columns a recommender uses for training.

aws personalize update-recommender \
--dataset-group-arn dataset group ARN \
--recommender-config "{\"trainingDataConfig\": {\"excludedDatasetColumns\": { \"datasetType\" : [ \"column1Name\", \"column2Name\"]}}}"

If you modify the columns used in training, Amazon Personalize automatically starts a full retraining of the models backing your recommender. While the update completes, you can still get recommendations from the recommender. The recommender uses the previous configuration until the update completes. To track the status of this update, use the latestRecommenderUpdate returned in the DescribeRecommender operation.

For more information about the different configurations you can change, see RecommenderConfig.

Updating a recommender (Amazon SDKs)

To update recommender with the Amazon, use the UpdateRecommender operation. Provide the Amazon Resource Name (ARN) for the recommender and specify the new configuration. The following code shows how to update the columns a recommender uses for training.

SDK for Python (Boto3)
import boto3

personalize = boto3.client('personalize')

update_recommender_response = personalize.update_recommender(
  recommenderArn = 'dataset group ARN',
  recommenderConfig = {
    "trainingDataConfig": {
      "excludedDatasetColumns": { 
        "datasetType": ["COLUMN_A", "COLUMN_B"]
      }
    }
  }     
)
SDK for JavaScript v3
// Get service clients and commands using ES6 syntax.
import { UpdateRecommenderCommand, PersonalizeClient } from
  "@aws-sdk/client-personalize";

// create personalizeClient
const personalizeClient = new PersonalizeClient({
  region: "REGION"
});

// set the request's parameters
export const updateRecommenderParam = {
  recommenderArn: "RECOMMENDER_ARN", /* required */
  recommenderConfig: {
    trainingDataConfig: {
      excludedDatasetColumns: {
        "DATASET_TYPE": ["COLUMN_A", "COLUMN_B"]   
      }
    }
  }
};

export const run = async () => {
  try {
    const response = await personalizeClient.send(new UpdateRecommenderCommand(updateRecommenderParam));
    console.log("Success", response);
    return response; // For unit tests.
  } catch (err) {
    console.log("Error", err);
  }
};
run();

If you modify the columns used in training in the excludedDatasetColumns of the recommenderConfig, Amazon Personalize automatically starts a full retraining of the models backing your recommender. While the update completes, you can still get recommendations from the recommender. The recommender uses the previous configuration until the update completes. To track the status of this update, use the latestRecommenderUpdate returned in the DescribeRecommender operation.

For more information about the different configurations you can change, see RecommenderConfig.