Best Approach for Recommender System Warm-Up in Gorse

[AnastasiiaKorneeva]

Hello @zhenghaoz,

First of all, thank you for your outstanding work on Gorse! We truly appreciate the effort and innovation behind this recommendation system.

We are developing a web application for event discovery in Barcelona, where Gorse plays a central role in providing personalized event recommendations based on user preferences and behavior.

Our Main Challenge: Recommender System Warm-Up
We are looking for the best approach to initialize recommendations when a user first joins our platform. Specifically, we aim to:

• Incorporate explicit user preferences (e.g., categories, districts, or budget selected during onboarding) into the initial recommendations.

• Ensure recommendations align with user choices while maintaining diversity within selected categories, even before sufficient interaction data is available.

We have reviewed Gorse’s feedback mechanism and ranking models, but we would greatly appreciate guidance on the best way to structure this warm-up phase efficiently. Are there any best practices, existing configurations, or specific model settings you would recommend for this use case?

Additionally, if there are relevant examples, documentation, or tuning strategies, we’d love to explore them. If you prefer to discuss this in more detail, we’d be happy to jump on a short call or connect in another way. Also, if you're interested, we’d love to explore ways to collaborate further or even get your insights on our implementation beyond this issue. No pressure—just putting it out there! 😃 If there’s a preferred way to contribute (e.g., documentation updates, sharing use cases, or commercial support), please let us know.

Looking forward to your insights, and thanks again for your incredible work!

Best regards,
Anastasiia
Flan app