基于两种数据挖掘算法的股骨颈预后评分分类

发布时间：2018-05-04 04:43

本文选题：股骨颈骨折 + 决策树C4.5　；参考：《太原理工大学》2017年硕士论文

【摘要】：股骨颈骨折手术预后质量评分(Harris评分)是骨科大夫极其关心的问题。随着病例的积累,我们希望通过病例信息找到影响Harris评分的重要因子,并使用这些影响因子对新患者的手术预后Harris评分类别进行预测。贝叶斯网络分类器是基于概率论与图论的分类网络,具有优良的分类功能,已广泛应用于数据挖掘、统计分析和人工智能等领域。决策树是一种基于信息增益理论的分类算法,是一种使用简单且应用面广泛的分类器。本文从数据挖掘的概述出发,首先,叙述了数据挖掘的基本内容及发展趋势;进一步介绍了决策树C4.5算法和贝叶斯网络分类器算法并提出了决策树C4.5算法的优化算法——决策树L-C4.5算法;最后,将决策树L-C4.5算法和贝叶斯网络分类器算法应用于股骨颈预后评分分类数据并成功搭建了优良的股骨颈手术预后评分的决策树和贝叶斯网络分类器。在此基础上,发现了Harris评分的重要影响因子分别为:BMI指数、骨折类型是否为Garden分型、是否存在糖尿病史、否是为侧位螺钉平行结构、骨折位置是否为三角结构及骨折类别。
[Abstract]:The prognosis quality score of femoral neck fracture (Harris score) is of great concern to orthopedic doctors. With the accumulation of cases, we hope to find out the important factors that affect the Harris score through the case information, and use these factors to predict the Harris score of the surgical prognosis of the new patients. Bayesian network classifier is a classification network based on probability theory and graph theory. It has excellent classification function and has been widely used in data mining, statistical analysis and artificial intelligence. Decision tree is a classification algorithm based on information gain theory. It is a simple and widely used classifier. This paper starts from the summary of data mining, first of all, describes the basic content and development trend of data mining; Furthermore, the decision tree C4.5 algorithm and Bayesian network classifier algorithm are introduced, and the decision tree L-C4.5 algorithm, which is the optimization algorithm of decision tree C4.5 algorithm, is proposed. The decision tree L-C4.5 algorithm and Bayesian network classifier algorithm are applied to the classification data of femoral neck prognosis score, and the excellent decision tree and Bayesian network classifier for the prognosis score of femoral neck surgery are successfully constructed. On this basis, it was found that the important influencing factors of Harris score were the Harris index, whether the fracture type was Garden classification, whether there was diabetes history, whether it was a lateral screw parallel structure, whether the fracture position was triangular structure and fracture type.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【参考文献】