基于节点输入策略贝叶斯网络的骨盆骨折分型研究

发布时间：2018-03-20 02:28

  本文选题：体表特征　切入点：贝叶斯网络　出处：《同济大学学报(自然科学版)》2017年08期 　论文类型：期刊论文

【摘要】：基于历史数据的统计和收集,选取骨盆骨折患者存在的18个体表特征,采用基于K2算法的贝叶斯网络方法挖掘各体表特征之间和骨盆骨折类型与体表特征间的相互关系;设计不同的节点输入策略,分析不同输入策略对算法性能的影响;基于骨盆稳定性将骨盆骨折分成A、B、C三种类型,分别找到与其直接相关的体表特征,作为判断骨盆骨折类型的依据.基于体表特征和骨盆骨折类型的分析结果,借助早期的观察及简单检查,对患者进行初步分型.
[Abstract]:Based on the statistics and collection of historical data, 18 body surface features of pelvic fracture patients were selected, and Bayesian network method based on K2 algorithm was used to mine the relationships between the surface features and the types of pelvic fractures and body surface features. Different node input strategies were designed to analyze the effect of different input strategies on the performance of the algorithm. Pelvic fractures were classified into three types based on pelvic stability. Based on the analysis results of surface features and pelvic fracture types, early observation and simple examination were used to classify the patients.
【作者单位】： 同济大学经济与管理学院;
【基金】：国家自然科学基金(71090404,71072026)
【分类号】：TP18
，

本文编号：1637116