基于产检数据聚类的妊娠合并症可视分析

Prenatal Data Cluster Based Pregnancy Complications via Visual Analysis

摘要:

妊娠合并症通常是指在尚未怀孕或妊娠期间由其他原因导致的疾病。很多情况下，孕妇在产检时各项检查结果表现正常，但却在之后的妊娠期间罹患妊娠合并症，所以单凭单项指标并不能很好的发现潜在的妊娠合并症。通过选取多组孕妇的多维产检指标数据进行聚类分析，在此基础上设计并开发了探索妊娠合并症潜在可能性的对比可视分析系统。案例分析表明，临床妇科医生能够通过可视化界面观察就诊孕妇妊娠期间的生理指标，查找与之相似病例，与传统就诊方式相比，提高发现其患有潜在妊娠合并症的概率已达到74.8%，平均误差率降低2.8%。

Abstract:

Pregnancy complication is the disease which happened in early or mid of pregnancy without certain reason. Some pregnant women maybe suffer from pregnancy complications after the maternity examination. However, those pregnant got a normal result in maternity examination, shows they have nice conditions. This indicated that the single indicator has lowly possible to detect the potential pregnancy complication. Through selecting a group of important dimensions in pregnancy test data for pregnant cluster analysis, a comparative visual analysis system for exploring and detecting the possibility of suffering pregnancy complication with the important dimensions of origin data was designed. The case study shows the system is able to help clinical gynecologists to detect the potential patient and find similar patients with exit cases. Besides, compared with the traditional medical treatment, the probability of finding potential pregnancy complications has increased to 74.8%, and the average error rate has decreased by 2.8%.