Title:
RISK PREDICTION METHOD AND APPARATUS BASED ON REINFORCEMENT LEARNING, AND DEVICE AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2023/108987
Kind Code:
A1
Abstract:
A risk prediction method and apparatus based on reinforcement learning, and a device and a storage medium, which relate to the field of artificial intelligence. The method comprises: receiving a risk prediction request regarding a target subject matter, wherein the risk prediction request comprises a prediction date (S201); acquiring a receipt date of the risk prediction request and target historical data of the target subject matter in N days before the receipt date (S202); performing feature extraction on the target historical data to obtain a target state feature corresponding to each of a plurality of preset feature dimensions (S203); and inputting the target state features into a first risk prediction model to obtain a risk value of the target subject matter on the prediction date (S204), wherein the first risk prediction model is a model that is obtained by means of optimizing a first Actor model on the basis of a first training set, a first Critic model or a second Critic model. By means of the method, the accuracy of risk prediction can be improved, thus facilitating risk decision making.
Inventors:
XIAO JING (CN)
GUO XIAO (CN)
WANG LEI (CN)
WANG YUAN (CN)
LIU YUNFENG (CN)
TAN TAO (CN)
CHEN YOUXIN (CN)
GUO XIAO (CN)
WANG LEI (CN)
WANG YUAN (CN)
LIU YUNFENG (CN)
TAN TAO (CN)
CHEN YOUXIN (CN)
Application Number:
PCT/CN2022/090029
Publication Date:
June 22, 2023
Filing Date:
April 28, 2022
Export Citation:
Assignee:
PING AN TECH SHENZHEN CO LTD (CN)
International Classes:
G06Q40/04; G06N3/04; G06Q10/04
Foreign References:
| CN114240656A | 2022-03-25 | |||
| CN112101520A | 2020-12-18 | |||
| CN112927085A | 2021-06-08 | |||
| CN107798608A | 2018-03-13 | |||
| CN112488826A | 2021-03-12 | |||
| CN110059896A | 2019-07-26 |
Attorney, Agent or Firm:
BEIJING CHENQUAN INTELLECTUAL PROPERTY LAW FIRM (CN)
Download PDF: