A new method for identification of traditional Chinese medicine constitution based on tongue features with machine learning

Technology and Health Care, Journal Year: 2024, Volume and Issue: unknown, P. 1 - 16

Published: May 25, 2024

BACKGROUND: The theory of Chinese medicine (TCM) constitution contributes to the optimisation individualised healthcare programmes. However, at present, TCM identification mainly relies on inefficient questionnaires with subjective bias. Efficient and accurate can play an important role in healthcare. OBJECTIVE: Building efficient model for identifying traditional constitutions using objective tongue features machine learning techniques. METHODS: DS01-A device was applied collect images extract features. We trained evaluated five models: Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), LightGBM (LGBM), CatBoost (CB). Among these, we selected best performance as base classifier constructing our heterogeneous ensemble model. Using various metrics, including classification accuracy, precision, recall, F1 score, area under curve (AUC), comprehensively evaluate performance. RESULTS: A total 1149 were obtained 45 extracted, forming dataset 1. RF, LGBM, CB learners RLC-Stacking. On 1, RLC-Stacking1 achieved accuracy 0.8122, outperforming individual classifiers. After feature selection, RLC-Stacking2 improved 0.8287, improvement 0.00165 compared RLC-Stacking1. exceeding 0.85 each type, indicating excellent CONCLUSION: study provides a reliable method rapid assist clinicians tailoring individualized medical treatments based personal types guide daily health care. information extracted from serves effective marker identification.

Language: Английский

Analysis of Traditional Chinese Medicine Health Data based on Data Mining: Constitution Classification and Personalized Wellness Recommendation System

Published: Dec. 27, 2024

Language: Английский