论文部分内容阅读
[目的]使用决策树中分类方法的CHAID算法模型和非条件Logistic回归模型,探讨新生儿低出生体重危险因素,并比较决策树的和Logistic模型间结果的异同。[方法]根据新生儿是否低出生体重分类。分别建立决策树CHAID模型和非条件Logistic模型。[结果]母亲孕周是新生儿低出生体重最主要的影响因素,在孕周小于37周的亚群中,多胎的新生儿发生低出生体重的风险更高。[结论在二分类结果变量的危险因素分析中,非条件Logistic模型和决策树模型都有较高的应用价值,两种方法可以相互结合,得到更为全面的结果。
[Objective] To explore the risk factors of low birth weight in neonates by using CHAID algorithm and non-conditional Logistic regression model in the classification of decision tree and to compare the similarities and differences between the results of decision tree and Logistic model. [Method] According to whether newborns have low birth weight classification. Establish decision tree CHAID model and unconditional Logistic model. [Results] Maternal gestational age was the most important factor affecting neonatal low birth weight. In sub-groups of gestational age less than 37 weeks, multiple births were at higher risk of having low birth weight. [Conclusion] In the analysis of risk factors of the two categorized result variables, both the non-conditional Logistic model and the decision tree model have higher application value, and the two methods can be combined to obtain more comprehensive results.