
Amazon Personalize now uses the latest streamed user interaction data for batch recommendations, improving recommendation quality without requiring model retraining.


<div><p>This article announces an improvement to Amazon Personalize's batch recommendations feature, which now incorporates the latest streamed user interaction data to enhance recommendation quality.</p><ul><li>Batch recommendations now use recently streamed events without requiring model retraining</li><li>Captures latest user interactions to better respond to behavior shifts</li><li>Previously only considered data up to last model retraining point</li><li>Simple implementation: create event tracker and run batch inference job</li><li>Improves personalization across websites, apps, ads, and emails</li></ul><p>Amazon Personalize now delivers more timely and accurate recommendations by leveraging real-time streamed data in batch processing workflows.</p></div>


Related articles