Chinese restaurant process

The Chinese Restaurant Process (CRP) is a metaphor for how data points (like customers) are grouped into clusters (like tables) without predefining the number of clusters. Imagine a restaurant with an infinite number of tables. Customers enter one by one: they sit at an existing table with a probability proportional to how many people are already there, or start a new table with a certain probability. This process favors popular clusters but allows new ones to form, making it useful for models where the number of groups is unknown beforehand.