Machine learning based intelligent auditing and anomaly detection method and system for financial vouchers
By locating the target attribution feature dimension of abnormal states in the machine learning model and redirecting it to a dedicated review routing channel, the problem of system resource congestion caused by missing data dimensions is solved, and efficient and intelligent review of financial vouchers is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BASECAMP DATA SERVICES GROUP CO LTD
- Filing Date
- 2026-04-22
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, the lack of data dimensions leads to abnormal output judgments from machine learning models, triggering default interception processes that continuously suspend processing threads, causing system computing resource congestion and affecting processing continuity.
By extracting multi-dimensional features from the financial vouchers to be processed, a basic voucher feature vector is generated. Inference calculations are then performed in the machine learning audit model to locate the target attribution feature dimension that causes the abnormal state. The default interception process is blocked, and the process is redirected to a dedicated review routing channel to capture supplementary supporting data for supplementary verification. Finally, the final audit status is output.
It enables automated capture and supplementation of external data under abnormal conditions, reduces the passive suspension of processing threads, alleviates computing resource congestion, and maintains the smoothness of computer processing concurrent audit tasks.
Smart Images
Figure CN122415246A_ABST