TPC-GCN: Deep learning for pulse pattern classification in traditional Chinese medicine

Abstract

Pulse diagnosis holds a pivotal role in traditional Chinese medicine (TCM) diagnostics, with pulse characteristics serving as one of the critical bases for its assessment. Accurate classification of these pulse pattern is paramount for the objectification of TCM. This study proposes an enhanced SMOTE approach to achieve data augmentation, followed by multi-domain feature extraction. Graph data structures with varying configurations are subsequently constructed to facilitate more profound insights into the intrinsic information within the data. Additionally, a multi-channel lightweight graph convolutional network (GCN) is devised. This network's core strategy lies in extracting diverse layers of information through parallel branches, integrating local structural information with adaptive weights, and employing attention-weighted fusion to improve classification accuracy and model robustness. The proposed network model achieved 91.68% accuracy, a mean F1 score of 92%, a mean recall rate of 92%, and a mean precision rate of 92% on the pulse dataset. The results demonstrate a marked improvement in pulse classification accuracy, validating the efficacy of this approach while offering new perspectives and methodologies for pulse signal classification research.