A method for predicting enterprise tax burden risk based on time sequence and association rule
By extracting the enterprise's personalized business cycle and combining it with association rules and time-series prediction models based on time decay weights, the high misjudgment rate and traceability difficulties in tax burden risk identification in existing technologies have been solved, achieving highly accurate tax burden risk prediction and traceability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-27
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, association rule algorithms and time series prediction algorithms cannot accurately match the business characteristics of enterprises in predicting corporate tax burden risks, resulting in insufficient accuracy and timeliness in tax burden risk identification and a high misjudgment rate.
By acquiring multi-source historical data of target enterprises, data preprocessing and periodic feature extraction are performed to establish a personalized business cycle list. A time decay weighted association rule algorithm is used to mine weighted business association rules. Historical tax burden rate time series data and business association rules are input into the time series prediction model for training to generate a tax burden rate prediction interval that integrates business association logic. The prediction is then double-validated by combining association matching and interval matching.
It has enabled accurate assessment and tracing of the causes of corporate tax burden risks, improved the accuracy of anomaly identification to over 90%, and increased tax audit efficiency by 60%.
Smart Images
Figure SMS_12