Intelligent identification method of traditional Chinese medicine five tastes based on electronic tongue and sensory joint modeling

By combining electronic tongue and sensory modeling, the problems of detection correlation, system missingness, chemical heterogeneity and model interpretability in the identification of the five flavors of traditional Chinese medicine were solved. The systematic identification and intensity classification of the five flavors of traditional Chinese medicine were realized, the identification accuracy and model credibility were improved, and the pharmacopoeia standardization and regulatory requirements were met.

CN122417467APending Publication Date: 2026-07-17SHANXI UNIV OF CHINESE MEDICINE
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202610456294.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-08
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Existing technologies for identifying the five flavors of traditional Chinese medicine suffer from several problems, including insufficient correlation between detection methods and taste perception, lack of a five-flavor identification system, difficulty in adapting modeling methods to chemical heterogeneity, lack of taste intensity grading methods, insufficient model interpretability, and the risk of data leakage during model validation.

Method used

We employ a joint modeling approach combining electronic tongue and sensory input. Through dual-channel data acquisition, feature fusion, unsupervised exploratory analysis, heterogeneous stacked ensemble learning, and SHAP interpretability analysis, we construct a model for identifying the five flavors and intensity grading of traditional Chinese medicine. This model is then used to perform correlation mapping in conjunction with traditional Chinese medicine theories of sexuality.

Benefits of technology

It has achieved systematic and objective identification and intensity classification of the five flavors of Chinese herbal medicines, improved the identification accuracy of a few categories, provided a highly interpretable model, ensured the generalization ability and credibility of the model, and met the requirements of pharmacopoeia standardization and regulatory approval.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122417467A_ABST
    Figure CN122417467A_ABST
Patent Text Reader

Abstract

本发明公开了一种基于电子舌和感官联合建模的中药五味智能辨别方法,属于中药分析技术领域。包括:中药样本采集与制备;对样本同步进行电子舌检测与人工感官评价以获取双渠道数据;对数据进行预处理与特征融合;采用异构堆叠集成学习框架构建五味识别模型,输出五味类别及置信度;采用数据驱动的自适应阈值确定方法与投票集成框架构建强度分级模型,输出强度分级结果及置信度;对模型进行SHAP可解释性分析并将结果与传统药性理论关联;对模型进行多维度评估与验证。本发明解决了现有技术中五味辨识体系缺失、小样本不平衡分类精度低、强度分级依赖经验阈值及模型可解释性不足的问题,实现了对中药五味的准确、客观、可解释的智能辨识与强度分级。
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Traditional Chinese medicine property identification method and system, computer equipment and storage medium

    CN114911977A