Verified75% confidenceFactExact time
The core optimization target of YouTube's recommendation algorithm has long been predicting user watch time.
3
Sources
75%
Confidence
Long-term
Relevance
8/11/2026
First Seen
Sources
如何避免意识形态极端化：保持理性思维的实用指南
Related Claims
UnverifiedYouTube在2012年将推荐机制从'点击率优化'切换至'观看时长优化'后观察到创作者行为趋同现象78% similarUnverified点击率始终是YouTube视频能否进入推荐池的第一道门槛，视频首先得被点击才有机会被留存率等后续指标评估68% similarUnverifiedYouTube has listed mass production and repetitive templates under the scope of its monetization policy review.67% similarUnverifiedBilibili's video recommendation algorithm comprehensively considers factors such as text relevance between video titles and search terms, historical click-through rates, interaction data (likes, coins, favorites, danmu), and the UP host account's weight.65% similarUnverifiedThe Agentic Loop mechanism enables the model to actively retrieve video content with purpose and decide through an iterative process whether more content needs to be examined65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/729581API
curl https://kongchang.com/api/v1/knowledge/claims/729581MCP
get_claim(id=729581)