一种面向项目实施过程的智能监管异常监督预警方法
By collecting multi-source heterogeneous data from the project monitoring server and constructing a finite state machine and a multi-task time series prediction model, combined with an adaptive state transition probability matrix and a manual review and feedback mechanism, the problems of process perception lag and data alignment in project monitoring were solved, achieving timeliness and accuracy of anomaly identification and reducing the risk of model drift.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-19
- Publication Date
- 2026-07-17
AI Technical Summary
Existing project supervision methods suffer from problems such as delayed process perception, unstable data time sequence alignment, difficulty in identifying abnormal process status, difficulty in distinguishing reasonable deviations from real anomalies, and model drift caused by error feedback.
By collecting multi-source heterogeneous data from the project monitoring server, a finite state machine is constructed to generate a unified time-series input vector. Combined with a multi-task time-series prediction model and multi-detector fusion, an adaptive state transition probability matrix and a manual review feedback mechanism are adopted to generate a structured chain of anomaly evidence.
It improves the timeliness, accuracy, stability and interpretability of anomaly identification during project implementation, reduces the impact of inconsistent data granularity across multiple systems on model input, and avoids self-reinforcing bias in error warning results.
Smart Images
Figure CN122222581B_ABST
Abstract
Citation Information
Patent Citations
Project progress monitoring and early warning method and device, computer equipment and storage medium
CN117132221A
Process execution supervision system and method suitable for project cost consultation project
CN120746507A