摘要
针对临床路径决策分析聚类算法中聚类效果依赖于样本数据分布且处理数据效率低的问题,提出基于均衡分配方法的模糊K-means算法的临床路径决策方法.该算法利用文字数字化处理与加权计算来建立数据格式统一且关键属性突出的样本特征值矩阵;利用基于均衡分配方法的模糊K-means算法对上述样本进行聚类分析,得到最终的聚类中心与聚类结果,以此辅助医生进行临床路径决策.采用ECLIPSE编程进行仿真,与传统模糊K-means算法和基于减法聚类的FCM算法相比,采用该算法的迭代时间分别降低了26%与70%,迭代次数分别减少了33%和82%,平均目标函数最小值分别减小了32%和28%.实验表明,该算法能够有效降低聚类效果对于样本数据分布的依赖,同时数据聚类效率与质量也有显著的提高.
To solve the problem of the dependence of clustering effect on the distribution of sample data and low data processing efficiency in clustering algorithm of clinical path decision analysis,we propose a fuzzy K-means algorithm clinical pathway decision method based on balanced allocation method.This algorithm establishes the matrix of sample eigenvalues with unified data format and obvious key attributes by using digital processing and weighted calculation of characters.The fuzzy K-means algorithm is used to do clustering analysis by the above samples,based on balanced distribution method,to assist doctors in clinical pathway decision with the help of the cluster center and cluster results.Compared with the traditional fuzzy K-means algorithm and FCM algorithm based on subtractive clustering,the iteration time of the algorithm reduces by26%and 70%respectively,and the iteration decreases by 33%and 82%respectively,and the average objective function minimum decreases by 32%and 28%respectively,by using emulation with ECLIPSE programming.The results indicate that the algorithm can effectively reduce the dependence of clustering effect on the distribution of sample data,and significantly improve the efficiency and quality of data clustering.
作者
郑帅
吕芳
ZHENG Shuai;LV Fang(Department of National Culture and Vocational Education,Liaoning National Normal College,Fuxin Liaoning 123000;Department of Neurology,Fuxin Center Hospital,Fuxin Liaoning 123000)
出处
《辽宁师专学报（自然科学版）》
2019年第3期81-88,共8页
Journal of Liaoning Normal College(Natural Science Edition)
关键词
聚类算法
均衡分配
加权
决策分析
clustering algorithm
balanced allocation
weighted
decision analysis
