User feature data calculation method and device

Abstract

The invention discloses a user feature data calculation method. The user feature data calculation method comprises the steps that activity data of a user is acquired, and a plurality of relevant users related with the user are determined according to the activity data of the user; multiple feature vectors between the user and the multiple relevant users are established according to the activity data of the user, and multiple relevant degrees between the user and the multiple relevant users are calculated according to the multiple feature vectors; multiple reference users are selected from the multiple relevant users according to the multiple relevant degrees; the feature data of the user is calculated according to the multiple reference users. According to the user feature data calculation method, mass data in the Internet is screened, by selecting the activity data of the multiple relevant users which are most relevant to the user to calculate the feature data, the calculation amount of the data is reduced, and the calculation speed is increased. The invention further discloses a user feature data calculation device.

Description

technical field [0001] The present application relates to the technical field of the Internet, in particular to a method and device for calculating user characteristic data. Background technique [0002] At present, user activity data is recorded on the Internet, and user characteristic data can be calculated through the user activity data, for example, user preferences, activity tracks, spending power, and the like. However, often due to the small amount of data of a single user recorded in the Internet (such as missing or very sparse data, etc.), it may lead to inaccurate calculation of user characteristic data, and the amount of all user activity data recorded in the entire Internet is too large. There are too many, and it is impossible to filter out the associated activity data used for the calculation of a certain user characteristic data, or the calculation speed is too slow due to the large amount of data on the Internet. Contents of the invention [0003] The purp...

AI Technical Summary

Problems solved by technology

However, often due to the small amount of data of a single user recorded in the Internet (such as missing or very sparse data, etc.), it may lead to inaccurate calculation of user characteristic data, and the amount of all user activity data recorded in the entire Internet is too large. There are too many, and it is impossible to filter out the associated activity data used for the calculation of a certain user characteristic data, or the calculation speed is too slow due to the large amount of data on the Internet, etc.

Images