基于案例推理的脑卒中康复训练方案设定模型

Setting Model of Rehabilitation Training Program for Stroke by Case-based Reasoning

针对康复训练方案制定过程具有非线性且难以用数学模型描述的问题,结合案例推理(CBR)技术,提出一种脑卒中康复训练方案设定模型.该模型根据患者病情,定义了一个新的相似度函数实现案例检索,并提出基于层次聚类的案例整合,不断丰富案例库,最终获得患者的康复训练方案.最后对100名偏瘫患者进行实验研究.实验结果表明:利用该模型制定康复训练方案对患者进行康复训练后,患者运动功能评分值基本控制在康复目标值范围内(>94分).与康复医师制定康复训练方案相比,患者运动功能评分值提高了0.49分,即该模型可有效地制定康复训练方案,使脑卒中患者早日实现生活自理.

For the problems that the process of drawing up a rehabilitation training program is nonlinear and is difficult to be described by accurate mathematical models, a model of rehabilitation training program for stroke is proposed, combining with the case-based reasoning （CBR）. According to the patients＇ conditions, a new similarity function is defined to implement case retrieval, a case integration method is presented based on hierarchical clustering for enriching case base, and then the rehabilitation training program is obtained. Experimental results on 100 patients show that using the rehabilitation training program based on the proposed model to guide rehabilitation training, the function score after training can be controlled in the range of rehabilitation target value （〉94）. Compared with rehabilitation physician, the patients＇ average score increases 0.49. It means that the proposed model can make the rehabilitation training program effectively, and make stroke patients look after themselves early.