脉象仪中的智能分类算法研究  

Study on Intelligent Classification Algorithms in Pulse Instruments

作　　者：田紫微 贾芸芳[1] TIAN Ziwei;JIA Yunfang(College of Electronic Information and Optical Engineering,Nankai University,Tianjin 300350,China)

机构地区：[1]南开大学电子信息与光学工程学院,天津300350

出　　处：《中国医疗设备》2024年第3期146-153,159,共9页China Medical Devices

基　　金：国家自然科学基金(62271269,61771260)。

摘　　要：脉诊借助脉搏探查人体气血循环状态,为疾病诊治、未病先知、养生保健等提供指导信息。然而,从微弱脉搏中提取脉象信息仍是脉象仪研发中的瓶颈问题。本文从脉搏信号特点、中医脉象分类及其面临的挑战等方面入手,综述了近年来利用基本机器学习(Machine Learning,ML)算法、神经网络算法以及集成学习算法建立脉象分类模型的相关研究,旨在通过比较不同ML算法和实验方案在脉象分类准确度上的表现,探讨基于ML算法建立脉象分类模型的可行性和有效性,以期为脉象仪研发提供参考。The pulse diagnosis detects the circulation status of Qi and blood in human body with the help of pulse so as to guide disease diagnosis and treatment,prediction of the disease and health care.However,the bottleneck problem in the development of pulse instruments is how to extract pulse information from weak pulse signals.This paper reviewed recent studies about used basic machine learning(ML)algorithms,neural network algorithms,and ensemble learning algorithms to build a pulse classification models,starting with issues such as the characteristics of pulse signals,classification of pulse conditions in traditional Chinese medicine,and the challenges they face.The purpose is to explore the feasibility and effectiveness of establishing models to classify pulse phases based on ML algorithms by comparing the performances of different ML algorithms and experimental protocols from the view point of the classification accuracy of different pulse phases,aims to provide a reference for the development of pulse instruments.

关 键 词：脉象仪 脉诊 脉象分类模型 机器学习算法 

分 类 号：R197.39[医药卫生—卫生事业管理] R241[医药卫生—公共卫生与预防医学]