摘要
为解决当前临床上缺乏对高血压疾病亚型及并发症发病模式分析的技术体系问题,本文提出了一种集成主元分析(PCA)、K-means聚类、Apriori频繁项挖掘等理论,对高血压患者群体差异因素下的并发症模式分析方法。首先,针对患者指标的多样性所带来的冗余干扰问题,利用PCA理论对指标数据进行降维及去冗余处理;其次,在获取指标数据主元成分的基础上,利用K-means算法实现患者的群体分析;最后,基于不同患者群体的并发症数据,利用Apriori算法实现并发症频繁模式分析。本文同时采用实际案例验证上述方法的有效性,以期为当前医疗大数据的分析与应用提供有效的解决思路与方案。
To solve the problem of lacking of the subtypes of hypertension and the pathogenesis of complications in current clinical analysis, an analysis model involving integrating principal components analysis (PCA), K-means clustering algorithm, and Apriori algorithm was proposed in this article. Firstly, according to the redundant interference problem caused by the diversity of the patients' clinical index, the PCA theory was used to reduce the dimension and the redundant relationship. Secondly, on the basis of obtaining the main component of the clinical index data, the K-means algorithm was used to conduct the patients' group analysis. Finally, the Apriori algorithm was used to analyze the frequent pattern of complications based on the complication data of different patients group. We used an example to verify efficacy of the above methods. The new analysis model of complications of hypertensive patients would provide an effective solution for the application of the current medical big data.
出处
《中国循证医学杂志》
CSCD
2017年第9期1100-1105,共6页
Chinese Journal of Evidence-based Medicine
基金
中央高校基本科研业务费专项资金-综合交叉项目(编号:xj2014108)
陕西省科技计划青年科技新星项目(编号:2015KJXX-06)