A rule and model fusion-based enterprise credit reporting risk identification system

The enterprise credit report risk identification system, which integrates rules and models, overcomes the limitations of single rule engines and single machine learning models, and achieves accurate identification and interpretable output of enterprise risks, adapting to different data quality scenarios and business needs.

CN122134365APending Publication Date: 2026-06-02LEXIANG DIGITAL CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
LEXIANG DIGITAL CO LTD
Filing Date
2026-02-10
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

In existing technologies, single rule engines struggle to identify implicit and non-linear risks. Single machine learning models experience a sharp drop in prediction accuracy when data is scarce in SMEs, and they cannot provide clear risk explanations, leading to high false negative rates and cold start problems. Data and decision-making are disconnected, resulting in insufficient robustness.

Method used

The enterprise credit report risk identification system adopts a rule-based and model-based fusion approach, which includes a multi-source data parsing module, an expert rule engine module, a machine learning model module, a fusion decision module, and an application output module. It forms a closed-loop identification system through signal interconnection, dynamically adjusts weights to compensate for data gaps, and provides interpretable risk identification results.

Benefits of technology

It achieves comprehensive coverage of both explicit and implicit risks, improves the accuracy and robustness of risk identification, avoids the cold start problem, meets the interpretability requirements of financial regulation, and supports multi-format data access and business scenario adaptation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122134365A_ABST
    Figure CN122134365A_ABST
Patent Text Reader

Abstract

The application discloses a kind of enterprise credit reporting risk identification systems based on rule and model fusion, it is related to risk identification system technical field.The present application includes multi-source data analysis module, expert rule engine module, machine learning model module, fusion decision module and application output module, the present application integrates five big core modules of multi-source data analysis, expert rule engine, machine learning model, fusion decision and application output, heterogeneous credit data is processed by OCR and NLP technology, in combination with red line blocking mechanism and integrated learning algorithm, based on the dynamic adjustment of fusion weight of data completeness, the depth of rule and model is realized Collaboration.The present application greatly improves the accuracy and robustness of risk identification, with efficient processing and strong interpretability, adapt to multi-format credit data and different scale enterprise scene, meet the needs of financial supervision and business operation, with wide application value.
Need to check novelty before this filing date? Find Prior Art