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

The tenets of Traditional Chinese Medicine (TCM) place considerable emphasis on the significance of individual physique classification and the necessity of personalised health management. This study proposes an innovative framework based on data mining, utilising a single deep learning model, the Long Short-Term Memory Network (LSTM), in conjunction with a number of optimisation techniques, for the classification of TCM health data and the generation of personalised health recommendations. The introduction of adaptive learning rate adjustment and regularisation strategies enabled the optimisation of the model training process and the improvement of the model's generalisation ability through the reduction of overfitting and the merging of components. Additionally, the feature selection technique was employed to extract the most representative physique-related features, and the classification effect was further optimised by the weighted loss function. The experimental results demonstrate that the proposed model exhibits superior performance in terms of accuracy in constitution classification and recommendation effectiveness when compared to the traditional method.