Abstract:

BACKGROUND: The collection of pulse wave is usually based on cable way, this measure brings inconvenience with what is not easy to move and expand. At the same time, the traditional pulse condition identification also restricted by artificial experience.
OBJECTIVE: In order to make up the inconvenience of the cable acquisition pulse wave and the limitation of traditional pulse condition identification by the doctor’s knowledge and experience, this paper put forward a wireless pulse acquisition and use the improved support vector machine algorithm to classify the pulse condition, in order to achieve the purposes of pulse condition wireless transmission and intelligent identification.
METHODS: Firstly, we used the HK2000B+ integrated pulse sensor to collect the pulse wave, then after the pulse signal disposal circuit processed, transferred it to the computer through the wireless transceiver module, then after the pretreatment by the matrix laboratory software, the time domain and frequency domain feature were picked up. Finally, the 160 cases of pulse were classified into four kinds of pulse condition by improved support vector machine algorithm. The kinds of pulse condition were flat, smooth, string and fine.
RESULTS AND CONCLUSION: The experimental results show that this method has the high classification rate, low complexity and higher classification accuracy in present classification methods. It has an important significance to realize the remote and reveal of the pulse-taking.