随机森林算法在膝骨关节炎穴位敏化研究中的应用初探  

Preliminary study on the application of random forest in acupoint sensitization of knee osteoarthritis

作　　者：徐桂兴 周玉梅[1] 孙宁 崔瑾 常小荣[3] 冀来喜 梁雷堃 罗廖君 刘晓佳 王丹[1] 赵凌[1] 蔡定均[1] 郑晖[1] 孙铭声 耿国燕 程建[5] 梁繁荣[1] XU Gui-xing;ZHOU Yu-mei;SUN Ning;CUI Jin;CHANG Xiao-rong;JI Lai-xi;LIANG Lei-kun;LUO Liao-jun;LIU Xiao-jia;WANG Dan;ZHAO Ling;CAI Ding-jun;ZHENG Hui;SUN Ming-sheng;GENG Guo-yan;CHENG Jian;LIANG Fan-rong(College of Acupuncture-Moxibustion and Tuina,Chengdu University of Traditional Chinese Medicine,Chengdu 610075,China;College of Acupuncture-Moxibustion and Tuina,Guizhou University of Traditional Chinese Medicine,Guiyang 550025 China;College of Acupuncture,Moxibustion&Tuina,Hunan University of Chinese Medicine,Changsha 410208,China;School of Acupuncture and Tuina,Shanxi University of Chinese Medicine,Taiyuan 030619,China;Lab of Computer Vision and Machine Intelligence,University of Electronic Science and Technology of China,Chengdu 610054,China)

机构地区：[1]成都中医药大学针灸推拿学院,成都610075 [2]贵州中医药大学针灸推拿学院,贵阳550025 [3]湖南中医药大学针灸推拿学院,长沙410208 [4]山西中医药大学针灸推拿学院,太原030619 [5]电子科技大学计算机视觉与机器智能实验室,成都610054

出　　处：《中华中医药杂志》2022年第10期6010-6013,共4页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家自然科学基金重大项目(No.81590950)。

摘　　要：目的:采用随机森林算法构建膝骨关节炎(KOA)患者犊鼻穴穴位敏化判别模型,以期为穴位是否敏化的判别及病情评估提供参考。方法:以916例KOA患者和222名健康受试者为研究对象,采集犊鼻穴体表温度、机械痛阈以及压痛阈的数据,运用SPSS 24.0软件进行数据分析,采用Python软件随机森林算法对犊鼻穴穴位敏化数据构建穴位敏化判别模型。结果:与健康受试者比较,KOA患者犊鼻穴温度显著升高、机械痛阈和压痛阈显著降低(P<0.01);穴位敏化判别模型能很好区分KOA患者与健康人(准确率77.85%),其中影像学3级(准确率82.76%)和临床3期(准确率88.37%)KOA患者区分效果最好。结论:穴位敏化判别模型可为穴位是否发生敏化的判别以及病情评估提供参考。Objective:To construct a sensitization discriminant model of Dubi(ST 35)acupoints in patients with knee osteoarthritis(KOA)by using random deep forest algorithm,in order to provide reference for the judgment of acupoint sensitization and disease assessment.Methods:Taking 916 KOA patients and 222 healthy subjects as the research objects,the data of body surface temperature,mechanical pain threshold and tenderness threshold of Dubi(ST 35)acupoint were collected.SPSS 24.0 software was used for data analysis.The random forest algorithm of Python software was used to construct the acupoint sensitization discriminant model on the sensitization data of Dubi acupoint.Results:Compared with healthy subjects,the temperature of Dubi(ST 35)acupoints in KOA patients increased,and the threshold of mechanical pain and tenderness decreased(P<0.01).The acupoint-sensitized discriminant model could well distinguish KOA patients from healthy people(accuracy rate 77.85%),including imaging grade 3(accuracy rate 82.76%)and clinical stage 3(accuracy rate 88.37%)KOA patients had the best distinction.Conclusion:The acupoint sensitization discriminant model can provide a reference for the judgment of acupoint sensitization and disease assessment.

关 键 词：膝骨关节炎 犊鼻穴 穴位敏化 随机森林 

分 类 号：R246.9[医药卫生—针灸推拿学]