Search⌘ K
AI Features

Problem Statement and Metrics

Explore how to formulate the problem statement and design key metrics for personalized feed ranking, focusing on maximizing user engagement. Understand offline and online metrics such as Click Through Rate and Normalized Cross-Entropy. Discover requirements for training, personalization, data freshness, scalability, and latency in large-scale ranking systems.

LinkedIn feed ranking

1. Problem statement

Design a personalized LinkedIn feed to maximize long-term user engagement. One way to measure engagement ...

LinkedIn feed ranking
LinkedIn feed ranking

On the LinkedIn feed, there are five major activity types:

  • Connections (A connects with B)
  • Informational
  • Profile
  • Opinion
  • Site-specific
  • Intuitively different activities have very different CTRClick Through Rate. This is important when we decide to build models and generate training data.
...