Title:
USER CLUSTERING AND FEATURE LEARNING METHOD AND DEVICE, AND COMPUTER READABLE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2021/093368
Kind Code:
A1
Abstract:
The present application provides a user clustering and feature learning solution. According to the solution, a clustering algorithm and an encoding and decoding model in a deep learning network are combined, transaction behavior sequences of users can be first determined on the basis of transaction behavior data of the users, and then the transaction behavior sequences of the users are encoded on the basis of an encoder of the deep learning network to generate deep features; while the users are clustered according to the deep features to obtain a clustering result, the deep features are decoded on the basis of a decoder of the deep learning network to obtain restored transaction behavior sequences; a learning target is determined according to the clustering result and a decoding result, and parameters of the encoder and the decoder of the deep learning network are iteratively adjusted according to the learning target, so that the deep learning network can be optimized while clustering is completed, and good deep features used for realizing clustering are obtained.
Inventors:
LI HUAISONG (CN)
PAN JIANMIN (CN)
PAN JIANMIN (CN)
Application Number:
PCT/CN2020/104002
Publication Date:
May 20, 2021
Filing Date:
July 24, 2020
Export Citation:
Assignee:
ALIPAY HANGZHOU INF TECH CO LTD (CN)
International Classes:
G06K9/62
Foreign References:
| CN111062416A | 2020-04-24 | |||
| CN109389166A | 2019-02-26 | |||
| CN108734338A | 2018-11-02 | |||
| CN105825269A | 2016-08-03 | |||
| US10068251B1 | 2018-09-04 |
Attorney, Agent or Firm:
BEIJING BESTIPR INTELLECTUAL PROPERTY LAW CORPORATION (CN)
Download PDF: