"Recommendation System Practice" 01_Good Recommendation System

Chapter 1 Good Recommendation System

1.1 What is a recommendation system

Realistic problem: information overload

The task of the recommendation system: contact users and information. (1) Help users find information that is valuable to them. (2) Let the information be displayed in front of users who are interested in it.

Ways to solve information overload:

1. Categories: Only cover a small number of popular websites.
2. Search engine: Users are required to actively provide accurate keywords to find information.
3. Recommendation system: It does not require users to provide clear needs, and models users' interests by analyzing users' historical behaviors, so as to actively help users recommend information that can meet their interests and needs.

1.2 Application of personalized recommendation system

1.3 Recommendation system evaluation

1.3.1 Recommended system experiment method

1. Offline experiment
2. User survey
3. Online experiment (AB test)

1.3.2 Evaluation Index

1. customer satisfaction
2. Forecast accuracy
3. Coverage
4. Diversity
5. Novelty
6. Surprise
7. Trust
8. real-time
9. Robustness
10. Business goals

1.3.3 Evaluation dimensions

1. User dimension
2. Item dimension
3. Time dimension