摘要
目的:尝试采用重复增量修剪法对457例卒中后吞咽功能患者的病历资料进行机器学习,探讨患者个人证候体征对显效率的影响。方法:从病历系统中提取符合纳入条件的舌三针治疗卒中后吞咽功能障碍患者的病例资料,运用重复增量修剪算法对所搜集的资料进行机器学习。结果:学习结果显示,吸烟与否、肥胖患者、血糖偏高、疾病分期等是对舌三针治疗卒中后吞咽功能障碍患者显效率影响最大的因素。结论:采用重复增量修剪法对卒中后吞咽功能患者显效率判断率高,可以协助临床医生做临床决策。
Objective: To explore the effect of personal syndrome on the markedly effective rate of 457 cases of postoperative patients with swallowing after stroke. Methods: The data of patients with swallowing dysfunction were extracted from the medical system,and the data were collected by PIPPER algorithm robotic learning. Results: The results showed that whether smoking,BMI index,disease staging,high blood surge were the most influential factors in the treatment of swallowing dysfunction. Conclusion: Using Repeated Incremental Pruning Algorithm for post-stroke swallowing patients with high rate of significant judgment,can help clinicians make clinical decisions.
引文
[1]KIDD D,LAWSON J,NESBITT R,et al.The natural history and clinical consequences of aspiration in acute stroke[J].QJM,1995,88(6):409-413
[2]曾莹莹,曾红梅,关中华.舌三针治疗假性球麻痹吞咽困难30例[J].福建中医药,2013,44(2):32-33
[3]陈熠路,陈明杨.针刺治疗脑卒中后吞咽障碍60例疗效观察[J].2014,36(11):66.
[4]BRETTLANTZ.机器学习与R语言[M].北京:机械工业出版社,2015:82-112.
[5]KURT HORNIK,CHRISTIAN BUCHTA,ACHIM ZEILEIS.Open-Source Machine Learning:R Meets Weka[J].Computational Statistics,2009,24(2):225-232.
[6]中国肥胖问题工作组.中国成人超重和肥胖症预防和控制指南(节录)[J].营养学报,2004,26(1):1-4
[7]中华神经科学会.各类脑血管疾病诊断要点[J].中华神经科杂志,1996,29(6):379.
[8]朱镛连.神经康复学[M].北京:人民军医出版社,2006:551.