User portrait is a tagged user model abstracted from the information of users' social attributes, living habits and consumption behaviors. The core work of building user portraits is to label users, which is to identify highly refined features through the analysis of user information.

Its role is generally inseparable from the following aspects:

Precision marketing, analysis of potential users of products, and SMS marketing for specific groups;

User statistics, such as the number of books purchased by China University TOP 10; ;

Data mining, building an intelligent recommendation system, using association rules to calculate what sports brands people who like red wine usually like, and using clustering algorithm to analyze the age distribution of people who like red wine;

Evaluate the effect, improve product operation and improve service quality. In fact, this is equivalent to market research and user research, quickly locate service groups and provide high-level services;

Personal tailor-made services or products, that is, personalized services for a certain group or even every user (personally, this is the current development trend and the mainstream of future consumption). For example, a company wants to launch a toy for children aged 5- 10. Through the analysis of users' portraits, it is found that the image = "Pleasant Goat" and the price range = "Medium" have the largest preference ratio, thus providing a very objective and effective decision-making basis for new products.