Abstract—Radial pulse signals have been found to be popular in ancient culture due to its simple, non invasive and effective approach for diagnosis. Traditional Chinese Medicine defines three pulse points on radial artery of both the hands for the assessment of health associated with internal body organs. An abnormal pulse pattern identified in the subjects of Prostate Enlargement is discussed with its characteristics. The feature vector for this abnormal pulse pattern and healthy pulse patterns are derived from power spectral density of the pulse signal. Dimensionality of the feature vector is reduced by ranking the frequency components according to their classification power. Binary pulse classifiers for the classification of such abnormal pulse pattern from the healthy pulse patterns are designed. Linear and Quadratic binary pulse classifiers give 89.0% and 89.2% classification accuracy for group of 5 and 4 initial ranked frequency components respectively.
Index Terms—Suppressed Dicrotic Notch, Traditional Chinese Medicine, Ranked Features, Binary Pulse Classifier.
Bhaskar Thakker is a research scholar in Instrument Design Development Centre (IDDC) at Indian Institute of Technology, Delhi, India (91-11-26596741; e-mail: thakkar_id@student.iitd.ac.in). Anoop Lal Vyas is a Professor in Instrument Design Development Centre (IDDC) at Indian Institute of Technology, Delhi, India (91-11-26591442; e-mail: alvyas@iddc.iitd.ac.in).
Cite: Bhaskar Thakker and Anoop Lal Vyas, "Suppressed Dicrotic Notch Pulse Classifier Design," International Journal of Machine Learning and Computing vol. 1, no. 2, pp. 148-153 , 2011.