关联规则优化的心脏疾病诱发因素检测算法  

Association Rule Optimization for Detection Algorithm of Factors Inducing Heart Disease

作　　者：毛颉[1] 王红玉[1] 

机构地区：[1]浙江工业职业技术学院,浙江绍兴312000

出　　处：《控制工程》2017年第6期1286-1290,共5页Control Engineering of China

摘　　要：针对现有的心脏疾病诊断系统耗时较多、昂贵且容易出错的问题,提出一种基于关联规则优化PSO-SVM的心脏疾病诱发因素检测算法。首先,利用关联规则挖掘算法选择疾病的特征,并对特征数据集进行训练;然后,PSO-SVM对训练集和测试集进行分类,并根据分类结果分析心脏疾病诱发因素;最后,在UCI克利夫兰数据集上以置信度作为指标的实验验证了提出的算法的有效性及可靠性。实验结果表明,相比其他两种较为先进的分类算法,提出的算法取得了更好的分类性能,为医生诊断和治疗心脏疾病提供了一个强有力的检测工具。The existing heart disease diagnose system takes too long time, costing much and making mistakes easily, for which a detection algorithm of factors inducing heart disease based on PSO-SVM optimized by association rules is proposed. Firstly, the association rule is used to select features from disease dataset so as to training features. Then, PSO-SVM is used to classify the training and testing sets and analyze factors of disease Finally, the effectiveness and reliability of the proposed algorithm has been verified by experiments on UCI Cleveland dataset. Analysis results show that the proposed algorithm has better classification efficiency than several advanced classification algorithms, which indicates that it supports a powerful detecting tool for doctors to diagnose and treat heart disease.

关 键 词：关联规则挖掘 心脏疾病 诱发因素检测 克利夫兰数据 支持向量机 粒子群优化 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]