Amazon Personalize is a fully managed machine learning service that uses your data to generate item recommendations for your users.
Recommendations
Amazon Personalize builds recommendation systems from your users’ interactions, such as views, clicks and purchases, with no ML to write.
How it works
- Import interaction data (who did what with which item) and optional user and item data.
- Choose a recommender or custom resource for the use case.
- Call it from the app to get recommendations or user segments in real time or in batch.
Exam cues
- If the question says personalized recommendations, “frequently bought together” or “more like this”, think Amazon Personalize.
- If the question says find users interested in an item, think Personalize user segments.
- If the question says trending items, think POPULAR_ITEMS; similar items, think RELATED_ITEMS; next best action, think PERSONALIZED_ACTIONS.
Remember
- Fully managed; uses your own interaction data.
- Recommendations for users, and user segments for items.
Trap: Grouping customers by similarity with no goal is clustering; recommending items to each user from their behavior is Amazon Personalize.
Da documentação da AWS (3)
- Amazon Personalize is a fully managed ML service that uses your data to generate item recommendations for users and user segments based on their affinity for items.
- Common Amazon Personalize uses include video recommendations such as "Top picks for you" and "More like X", and ecommerce recommendations such as "Frequently bought together" and "Customers who viewed X also viewed".
- Amazon Personalize recipes are algorithms prepared for specific use cases: USER_PERSONALIZATION (recommend items for a user), PERSONALIZED_RANKING (re-rank a list), POPULAR_ITEMS (trending or popular), RELATED_ITEMS (similar items), PERSONALIZED_ACTIONS (next best action) and USER_SEGMENTATION (user segments).
Flashcards
Nada para revisar aqui agora. Escolha outros tópicos ou outro filtro.
·
Todos os 5 flashcards em lista
An online store wants "Customers who viewed this also viewed" suggestions without building ML models. Which AWS service fits?
Amazon Personalize.
A streaming service wants a "Top picks for you" row per viewer. Which AWS service fits?
Amazon Personalize, which generates personalized recommendations from interaction data.
A marketing team wants to find the users most likely to be interested in a new product line. Which Amazon Personalize capability fits?
User segmentation, which groups users by their affinity for items or item attributes.
Does a company need ML expertise to add recommendations with Amazon Personalize?
No. It's fully managed; you provide interaction and item data, and it trains and serves the recommenders.
Which AWS service can recommend products to shoppers based on their affinity for items?
Amazon Personalize is a fully managed ML service that generates item recommendations using your data.