Intelligent advertisement recommendation system and method based on user portraits
By constructing a noise filtering mechanism based on the timestamp sequence of interactive events and a dynamic label weight update method, the problem of noise pollution in the intelligent advertising recommendation system is solved, achieving accurate identification of user interests and a balance between the diversity of recommendation results.
CN121597913APending Publication Date: 2026-03-03SHENZHEN HOUSELAI TECH CO LTD
View PDF 0 Cites 0 Cited by
Patent Information
- Application Number
- CN202511830797.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-06
- Publication Date
- 2026-03-03
Smart Images
Figure CN121597913A_ABST
Abstract
The invention relates to the technical field of intelligent advertisement recommendation, and discloses an intelligent advertisement recommendation system and method based on a user portrait. The method comprises the steps that a first influence coefficient is calculated based on the effective stay duration of a user browsing page, a second influence coefficient is calculated based on a time interval, the first influence coefficient and the second influence coefficient are multiplied and then accumulated to a first label weight vector, and a second label weight vector is obtained; and obtaining a label coverage vector of a candidate advertisement, performing dot product operation on the second label weight vector and the label coverage vector to obtain a matching score of the candidate advertisement, and generating an advertisement recommendation list according to the matching score. According to the method, the recommendation result is effectively prevented from collapsing towards a single label while the recommendation correlation is ensured, and the balance of accuracy and diversity is realized.
Need to check novelty before this filing date? Find Prior Art