A method and system for constructing a database of children's snack additive consumption
By using a dynamic identification algorithm based on feature weights and time series stability, the problem of inaccurate identification of abnormal samples in the construction of a children's snack additive database in traditional methods has been solved, thus achieving the construction of a high-quality database and improving the accuracy of risk assessment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- TIANJIN CENT FOR DISEASE CONTROL & PREVENTION
- Filing Date
- 2026-06-11
- Publication Date
- 2026-07-17
AI Technical Summary
When constructing a database of children's snack additive consumption, existing technologies cannot distinguish normal fluctuations in children's snack intake behavior using traditional dynamic threshold outlier identification algorithms. This results in the erroneous removal of a large number of real and valuable extreme consumption samples, leading to overly optimistic risk assessment results and creating potential safety hazards.
An outlier dynamic identification algorithm based on feature weights is adopted, combined with time series stability index, to dynamically adjust weights, identify and delete abnormal samples, and build a high-quality database of children's snack additive consumption.
It enables intelligent identification of anomalous samples, ensuring that high-impact features receive appropriate attention in risk calculation, keenly capturing potential risks, and improving the accuracy and security of risk assessment.
Smart Images
Figure CN122417455A_ABST