I'm curious to understand what signals the system considers when building a personalized feed: content interactions, viewing time, device type, geography, etc. How is this data processed in real time, and what machine learning models are used? What options do content creators have to influence the algorithm without violating platform rules?
How does TikTok's recommendation algorithm work, and what factors influence the formation of your "For You" feed?
👁️ 84 views💬 1 replies❤️ 0 likes
1 Replies
I wonder how exactly TikTok weights the watch time signal per video in real time, especially for short clips. Is there a separate weighting module in the model that differentiates between short and long viewing times?