HRNN-Metadata recipe (legacy) - Amazon Personalize

HRNN-Metadata recipe (legacy)

Note

Legacy HRNN recipes are no longer available. This documentation is for reference purposes.

We recommend using the aws-user-personalizaton (User-Personalization) recipe over the legacy HRNN recipes. User-Personalization improves upon and unifies the functionality offered by the HRNN recipes. For more information, see User-Personalization recipe.

The HRNN-Metadata recipe predicts the items that a user will interact with. It is similar to the HRNN recipe, with additional features derived from contextual, user, and item metadata (from Interactions, Users, and Items datasets, respectively). HRNN-Metadata provides accuracy benefits over non-metadata models when high quality metadata is available. Using this recipe might require longer training times.

The HRNN-Metadata recipe has the following properties:

  • Nameaws-hrnn-metadata

  • Recipe Amazon Resource Name (ARN)arn:aws:personalize:::recipe/aws-hrnn-metadata

  • Algorithm ARNarn:aws:personalize:::algorithm/aws-hrnn-metadata

  • Feature transformation ARNarn:aws:personalize:::feature-transformation/featurize_metadata

  • Recipe typeUSER_PERSONALIZATION

The following table describes the hyperparameters for the HRNN-Metadata recipe. A hyperparameter is an algorithm parameter that you can adjust to improve model performance. Algorithm hyperparameters control how the model performs. Featurization hyperparameters control how to filter the data to use in training. The process of choosing the best value for a hyperparameter is called hyperparameter optimization (HPO). For more information, see Hyperparameters and HPO.

The table also provides the following information for each hyperparameter:

  • Range: [lower bound, upper bound]

  • Value type: Integer, Continuous (float), Categorical (Boolean, list, string)

  • HPO tunable: Can the parameter participate in hyperparameter optimization (HPO)?

Name Description
Algorithm Hyperparameters
hidden_dimension

The number of hidden variables used in the model. Hidden variables recreate users' purchase history and item statistics to generate ranking scores. Specify a greater number of hidden dimensions when your Item interactions dataset includes more complicated patterns. Using more hidden dimensions requires a larger dataset and more time to process. To decide on the optimal value, use HPO. To use HPO, set performHPO to true when you call CreateSolution and CreateSolutionVersion operations.

Default value: 43

Range: [32, 256]

Value type: Integer

HPO tunable: Yes

bptt

Determines whether to use the back-propagation through time technique. Back-propagation through time is a technique that updates weights in recurrent neural network-based algorithms. Use bptt for long-term credits to connect delayed rewards to early events. For example, a delayed reward can be a purchase made after several clicks. An early event can be an initial click. Even within the same event types, such as a click, it’s a good idea to consider long-term effects and maximize the total rewards. To consider long-term effects, use larger bptt values. Using a larger bptt value requires larger datasets and more time to process.

Default value: 32

Range: [2, 32]

Value type: Integer

HPO tunable: Yes

recency_mask

Determines whether the model should consider the latest popularity trends in the Item interactions dataset. Latest popularity trends might include sudden changes in the underlying patterns of interaction events. To train a model that places more weight on recent events, set recency_mask to true. To train a model that equally weighs all past interactions, set recency_mask to false. To get good recommendations using an equal weight, you might need a larger training dataset.

Default value: True

Range: True or False

Value type: Boolean

HPO tunable: Yes

Featurization hyperparameters
min_user_history_length_percentile

The minimum percentile of user history lengths to include in model training. History length is the total amount of data about a user. Use min_user_history_length_percentile to exclude a percentage of users with short history lengths. Users with a short history often show patterns based on item popularity instead of the user's personal needs or wants. Removing them can train models with more focus on underlying patterns in your data. Choose an appropriate value after you review user history lengths, using a histogram or similar tool. We recommend setting a value that retains the majority of users, but removes the edge cases.

For example, setting min__user_history_length_percentile to 0.05 and max_user_history_length_percentile to 0.95 includes all users except those with history lengths at the bottom or top 5%.

Default value: 0.0

Range: [0.0, 1.0]

Value type: Float

HPO tunable: No

max_user_history_length_percentile

The maximum percentile of user history lengths to include in model training. History length is the total amount of data about a user. Use max_user_history_length_percentile to exclude a percentage of users with long history lengths because data for these users tend to contain noise. For example, a robot might have a long list of automated interactions. Removing these users limits noise in training. Choose an appropriate value after you review user history lengths using a histogram or similar tool. We recommend setting a value that retains the majority of users but removes the edge cases.

For example, setting min__user_history_length_percentile to 0.05 and max_user_history_length_percentile to 0.95 includes all users except those with history lengths at the bottom or top 5%.

Default value: 0.99

Range: [0.0, 1.0]

Value type: Float

HPO tunable: No