Amazon Personalize
Developer Guide

Using Predefined Recipes

Amazon Personalize provides predefined recipes to train a model, which enables you to create a personalization system without needing machine learning experience.

These recipes use predefined attributes of your data, predefined feature transformations, predefined algorithms, and initial parameters for the algorithms. You can override many of these parameters when creating a solution.

Amazon Personalize can automatically choose the most appropriate HRNN recipe (performAutoML) based on its analysis of the input data. Alternatively, you can choose a specific recipe based on your experience. Each recipe has its own use case, as described below, and you should select the recipe that best fits your needs.

You can see a list of available recipes in the Amazon Personalize console or list the recipes by calling the ListRecipes API. You can get information about a specific recipe by calling the DescribeRecipe API.

Amazon Personalize provides three types of recipes. Besides behavioral differences, each type has different requirements for getting recommendations, as shown in the following table.

Recipe type API userId itemId inputList
USER_PERSONALIZATION GetRecommendations required optional NA
PERSONALIZED_RANKING GetPersonalizedRanking required NA list of itemId's
RELATED_ITEMS GetRecommendations not used required NA

Predefined Recipes

Amazon Personalize provides the following predefined recipes listed by recipe type.

USER_PERSONALIZATION recipes

Predicts items a user will interact with.

HRNN^

A hierarchical recurrent neural network, which can model the temporal order of user-item interactions.

Recommended when user behavior is changing with time (the evolving intent problem).

HRNN-Metadata^*

HRNN with additional features derived from contextual metadata (Interactions dataset), along with user and item metadata (Users and Items datasets). The training data must include metadata in at least one of the datasets.

Performs better than non-metadata models when high quality metadata is available. Can involve longer training times.

HRNN-Coldstart^*

Similar to HRNN-metadata with personalized exploration of new items. The dataset group supplying the training data must include an Items dataset.

Recommended when frequently adding new items to a catalog and you want the items to immediately show up in recommendations.

Popularity-Count

Calculates popularity of items based on a count of events against that item in the user-item interactions dataset.

Use as a baseline to compare other user-personalization recipes.

PERSONALIZED_RANKING recipes

Personalizes results.

Personalized-Ranking

Use this recipe when you’re personalizing the results for your users, such as, personalized reranking of search results or curated lists.

RELATED_ITEMS recipes

Predicts items similar to a given item.

SIMS

Item-to-item similarities (SIMS) generates items similar to a given item based on co-occurrence of the item in user history in the user-item interaction dataset. In the absence of sufficient user behavior data for an item, or if the specified item ID is not found, the algorithm returns popular items as recommendations.

Use for improving item discoverability and in detail pages. Provides fast performance.

^ Amazon Personalize examines these recipes when performing AutoML.

* Metadata models: These recipes train on both interaction data and metadata. For more information, see Datasets and Schemas.

On this page: