Tradition Chinese Medicine (TCM) pulse analysis

• Main Author: Wee, Melvin Pei Shan
• Other Authors: Ser Wee
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/74893
• Institution: Nanyang Technological University

Pulse diagnosis is a commonly used diagnosis method in the Traditional Chinese Medicine (TCM) diagnosis. It is an effective tool for the assessment of the patient’s health condition. However, the accuracy of diagnosis result is heavily dependent on the TCM practitioner’s skills and experience. Thus, much research on quantitative methods is needed to be done to better aid the analysis of the TCM pulses. This project aims to characterize the Xi pulse (细脉) and the Xian pulse (弦脉) signal processing techniques. The project proposes the use of Discrete Wavelet Transform (DWT) to decompose a total of 60 pulse signals to extract time-frequency domain information as features for the pulse classification. The Fisher’s Ratio value of the features is computed and the features are ranked according to their Fisher’s Ratio value. The 152th detail coefficient value of the 2nd level decomposition and 183th detail coefficient value of the 1st level decomposition is chosen as the features for classification. The classification of the pulse signal is done using 2 types of machine learning algorithm, the Support Vector Machine (SVM) and the k-nearest neighbors (k-NN). A 10-fold cross validation is used to give a single estimation of the classification performace of the classifier. After evaluating the classification accuracy of the 2 types of classifiers, it can be concluded that both the classifiers have give a high classification accuracy of about 90.5% when using the 152th detail coefficient value of the 2nd level decomposition and 183th detail coefficient value of the 1st level decomposition of the Xi pulse and the Xian pulse as features.