Title:
PUSH MODEL TRAINING METHOD, DATA PUSH METHOD AND APPARATUS, AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2022/198752
Kind Code:
A1
Abstract:
Embodiments of the present application provide a push model training method, comprising: obtaining multiple pieces of user data of multiple sample users, and preprocessing the multiple pieces of user data to obtain a plurality of sample user clusters; inputting multiple pieces of user data of each sample user into a model to be trained, and outputting a first sample preference of each piece of item type data corresponding to each sample user; encoding the first sample preference corresponding to each piece of item type data to obtain corresponding first sample preference encoded data, and determining sample preference output encoded data therefrom; obtaining sample preference label encoded data corresponding to the preference sample item type data of each sample user; and according to a cost function, comparing the sample preference output encoded data of each sample user with the sample preference label encoded data to adjust model parameters of said model to obtain an item data recommendation model. The present application effectively improves efficiency and accuracy of item type data pushing of an optimized model.
Inventors:
CHENG LEI (CN)
Application Number:
PCT/CN2021/091176
Publication Date:
September 29, 2022
Filing Date:
April 29, 2021
Export Citation:
Assignee:
PING AN TECH SHENZHEN CO LTD (CN)
International Classes:
G06F30/27
Foreign References:
| CN107801096A | 2018-03-13 | |||
| CN102376063A | 2012-03-14 | |||
| CN103336793A | 2013-10-02 | |||
| CN105205081A | 2015-12-30 | |||
| US20190197013A1 | 2019-06-27 |
Attorney, Agent or Firm:
INTELLECPRO CHINA LIMITED (CN)
Download PDF:
Previous Patent: RAPID FACIAL DETECTION METHOD BASED ON MULTI-LAYER PREPROCESSING
Next Patent: DISPLAY PANEL AND DISPLAY APPARATUS
Next Patent: DISPLAY PANEL AND DISPLAY APPARATUS