世界中西医结合杂志
世界中西醫結閤雜誌
세계중서의결합잡지
World Journal of Integrated Traditional and Western Medicine
2015年
10期
1430-1433,1437
,共5页
汪南玥%于友华%刘佳%李同达%薛利媛%黄大威%江柏华%王佳%李彩芬
汪南玥%于友華%劉佳%李同達%薛利媛%黃大威%江柏華%王佳%李綵芬
왕남모%우우화%류가%리동체%설리원%황대위%강백화%왕가%리채분
滑脉%脉诊信息%脂肪肝%慢性阻塞性肺疾病
滑脈%脈診信息%脂肪肝%慢性阻塞性肺疾病
활맥%맥진신식%지방간%만성조새성폐질병
Rolling Pulse%Pulse Diagnostic Information%Fatty Liver%Chronic Obstructive Pulmonary Disease
目的：研究同为滑脉的两种不同疾病患者脉诊信息的变化特征，为传统脉象可再分提供依据。方法采用“中医脉诊信息采集系统”分别采集同为滑脉慢性阻塞性肺病（COPD）患者与脂肪肝（FLD）患者各50例的左右手寸、关、尺部位的脉诊信息特征，应用 MATLAB 脉诊信息 GUI 分析程序进行脉诊信息预处理、193个参数提取和主成分分析、LS 识别、Lasso3种不同分析方法的分类判决，分析2类人群脉诊信息之间的差异以及各自的信息特征。结果同为滑脉的2种疾病患者的脉诊信息之间存在很明显的差异，该差异在采用3种不同的数学方法分别进行统计的时具有较强的一致性。2种疾病患者的脉诊信息分析中，无监督学习方法分类识别的准确概率为67%，有监督学习的准确概率达79%-92%。2种疾病患者的脉诊信息之间的差异呈现规律性和部位特异性。结论滑脉在不同疾病状态下脉诊信息之间具有显著差异，可以进行再分类。
目的：研究同為滑脈的兩種不同疾病患者脈診信息的變化特徵，為傳統脈象可再分提供依據。方法採用“中醫脈診信息採集繫統”分彆採集同為滑脈慢性阻塞性肺病（COPD）患者與脂肪肝（FLD）患者各50例的左右手吋、關、呎部位的脈診信息特徵，應用 MATLAB 脈診信息 GUI 分析程序進行脈診信息預處理、193箇參數提取和主成分分析、LS 識彆、Lasso3種不同分析方法的分類判決，分析2類人群脈診信息之間的差異以及各自的信息特徵。結果同為滑脈的2種疾病患者的脈診信息之間存在很明顯的差異，該差異在採用3種不同的數學方法分彆進行統計的時具有較彊的一緻性。2種疾病患者的脈診信息分析中，無鑑督學習方法分類識彆的準確概率為67%，有鑑督學習的準確概率達79%-92%。2種疾病患者的脈診信息之間的差異呈現規律性和部位特異性。結論滑脈在不同疾病狀態下脈診信息之間具有顯著差異，可以進行再分類。
목적：연구동위활맥적량충불동질병환자맥진신식적변화특정，위전통맥상가재분제공의거。방법채용“중의맥진신식채집계통”분별채집동위활맥만성조새성폐병（COPD）환자여지방간（FLD）환자각50례적좌우수촌、관、척부위적맥진신식특정，응용 MATLAB 맥진신식 GUI 분석정서진행맥진신식예처리、193개삼수제취화주성분분석、LS 식별、Lasso3충불동분석방법적분류판결，분석2류인군맥진신식지간적차이이급각자적신식특정。결과동위활맥적2충질병환자적맥진신식지간존재흔명현적차이，해차이재채용3충불동적수학방법분별진행통계적시구유교강적일치성。2충질병환자적맥진신식분석중，무감독학습방법분류식별적준학개솔위67%，유감독학습적준학개솔체79%-92%。2충질병환자적맥진신식지간적차이정현규률성화부위특이성。결론활맥재불동질병상태하맥진신식지간구유현저차이，가이진행재분류。
Objective To study the change characteristics of rolling pulse information in different diseases so as to provide evidence for the re - categorization of traditional pulse condition. Methods TCM pulse diagnosis information collection system was adopted to collect the pulse information characteristics on cun,guan and chi regions of hands in each of the 50 patients of chronic obstructive pulmonary disease (COPD)and fatty liver disease(FLD)with rolling pulse. GUI analyzing process of MATLAB pulse diagnostic information was used for the preprocessing of pulse diagnostic information,extraction of 193 parameters and cauterization with different methods such as principal component analysis,LS identification and Lasso3. The differences in pulse diagnostic information and individual information characteristics were analyzed between the two groups. Results The differences were very apparent in the pulse diagnostic information between the two groups of rolling pulse. These differences were high in consistency with 3 different mathematics methods adopted. In the analysis,the accurate probability of categorization identification was 67% with unsupervised learning method and was 79% to 92% with supervised learning method. The differences between the two groups presented regularity and regional specificity. Conclusion In the condition of different diseases,thep-ulse diagnostic information of rolling pulse is significantly different and can be re - categorized.
