Diabetes feature selection method and system based on optimized dogvessel algorithm
By introducing a Logistic chaotic mapping mechanism and an adaptive mechanism to optimize the tunic algorithm, the problems of local optima and high computational complexity in feature selection are solved, achieving efficient and stable feature selection and improving the accuracy of the diabetes diagnostic model.
Patent Information
- Application Number
- CN202511796040.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-02
- Publication Date
- 2026-03-06
AI Technical Summary
Existing feature selection methods are prone to getting trapped in local optima in high-dimensional feature spaces, resulting in low search efficiency, unstable results, and high computational complexity, making it difficult to meet the needs of practical applications.
A Logistic chaotic mapping mechanism is introduced for initializing the tunic algorithm. The leader position is calculated by combining adaptive weights and preset perturbation terms, and the follower position is calculated by preset following factors and perturbation terms. Feature selection is optimized by binarization mapping and fitness calculation.
It improves the accuracy, stability, and repeatability of feature selection, enhances search efficiency, reduces redundant features, and improves the accuracy of diabetes diagnostic models.
Smart Images

Figure CN121617652A_ABST