A government affair data monitoring analysis processing method and system based on big data
By constructing a multidimensional government affairs time series tensor and an improved TimesFM model, combined with causal discovery algorithms and backdoor criteria, the high false alarm rate and computational resource consumption of traditional government affairs data anomaly detection methods are solved, achieving efficient anomaly detection and root cause localization in complex government affairs systems.
Patent Information
- Application Number
- CN202610546053.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-23
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional methods for detecting anomalies in government data consume enormous computational resources in multi-source heterogeneous government systems, have a high false alarm rate, cannot identify the dynamic causal relationships behind the data, and lack characterization of the underlying data generation mechanisms, resulting in an inability to accurately locate the root cause and severely restricting response efficiency.
This paper adopts a big data-based government data monitoring and analysis method. By collecting data from multiple heterogeneous government systems, it performs spatiotemporal granular alignment and missing value imputation to construct a multidimensional government time series tensor. It uses an improved TimesFM time series model to extract deep time series dependency features, combines causal discovery algorithms and backdoor criteria to identify confounding factors, constructs a deconfounded causal structure equation model, generates a dynamic anomaly threshold curve, and accurately distinguishes between normal policy fluctuations and real business violations.
It significantly reduces the false alarm rate of anomaly detection, improves the model's prediction accuracy and response efficiency, and can accurately distinguish between normal policy fluctuations and real business violations in complex intervention environments, dynamically deducing the expected baseline trajectory after removing the impact of policy intervention.
Smart Images
Figure CN122414550A_ABST