一种物料推荐方法、装置、设备及计算机可读存储介质
By introducing interest-matching features and a deep recommendation model, the shortcomings of existing recommendation systems in predicting user click-through rates are addressed, resulting in more accurate and diverse material recommendations and improved user click-through rates.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- MICRO DREAM TECHTRONIC NETWORK TECH CHINACO
- Filing Date
- 2022-12-16
- Publication Date
- 2026-07-17
AI Technical Summary
Existing recommendation systems, when predicting user click-through rates for materials, rely on Cartesian products and one-sided business cross-feature construction methods that are unsuitable for all features, resulting in inaccurate predictions of user click-through rates and a lack of generalization.
Using interest-matching features, including the number of matches and the matching ratio, a deep recommendation model is used to combine user features and material features to predict the click-through rate (CTR) of candidate materials and recommend materials to target users based on the predicted values.
It improved the accuracy and diversity of material recommendations, achieving more precise material recommendations and increasing user click-through rates.
Smart Images
Figure CN115840896B_ABST