A large model-based risk prevention and control method and electronic equipment

By cleaning and cross-validating multi-source business data through a large model and a set of preset risk control rules, a comprehensive risk rating is generated, which solves the problems of data interference and misjudgment in enterprise risk prevention and control, and achieves efficient and accurate risk identification and assessment.

CN121766787BActive Publication Date: 2026-06-12CHIA TAI TIANQING PHARMA GRP CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
CHIA TAI TIANQING PHARMA GRP CO LTD
Filing Date
2026-03-03
Publication Date
2026-06-12

AI Technical Summary

Technical Problem

Existing technologies are insufficient for efficient and accurate risk identification and assessment in enterprise risk control, especially when dealing with multi-source heterogeneous data and unstructured data, which pose risks of data interference and misjudgment.

Method used

By employing a large model combined with a pre-set set of risk control rules, multi-source business data is cleaned and its compliance is verified to generate standard business data. Business association keys are used to generate risk tasks to be assessed. Through cross-validation of structured and unstructured data, a comprehensive risk rating is generated, and the handling actions are determined based on the rating.

Benefits of technology

It improves the efficiency and accuracy of risk prevention and control, effectively identifies and assesses potential risks in complex business scenarios, reduces data interference, and enhances automation and decision-making accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121766787B_ABST
    Figure CN121766787B_ABST
Patent Text Reader

Abstract

The application discloses a risk prevention and control method based on a large model and electronic equipment, comprising: invalid data filtering and compliance verification are performed on multi-source business data related to an initiated financial reporting request, and standard business data is obtained through data cleaning; a to-be-evaluated risk task is generated according to a business association key matched by the standard business data; rule verification is performed on structured business data in the to-be-evaluated risk task according to a risk control rule determined according to the business association key, and a structured risk clue is generated; a first risk clue is determined in unstructured business data of the to-be-evaluated risk task by using a large model, and a corresponding composite risk control rule is determined according to an actual risk type to which the first risk clue belongs; a second risk clue corresponding to the actual risk type is extracted from the structured risk clue, and the composite risk control rule, the first risk clue and the second risk clue are analyzed by means of the large model to obtain a risk rating; and a disposal operation corresponding to the to-be-evaluated risk task is determined according to the risk rating.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • CN120278534A

  • CN120851032A