【作者】 李大鹏； 王惠文；

【Author】 Dapeng Li and Huiwen Wang School of Economics and Management, BeiHang University, Beijing 100083

【机构】 北京航空航天大学经济管理学院；

【摘要】 偏最小二乘logistic回归模型是一种新型的多元分析方法,它在自变量之间存在强多重共线性、或者当样本量偏小以及样本中有缺失值的情况下,可以较好地解决普通logistic回归模型的计算结果不稳定的问题。本文利用偏最小二乘logistic回归算法,根据鄱阳湖地区,1953～1998年观测的水文数据,分析各月连续最大五天降水量和长江各月最大流量对鄱阳湖洪涝灾害的影响,建立了预测洪涝灾害程度的发生概率的判别模型。研究结果表明,偏最小二乘logistic回归模型在相关领域的研究中具有很好的适用性。更多还原

【Abstract】 PLS logistic regression is a novel approach of multiple analysis proposed hy V. E. Vinzi and M. Tenenhuas in 2001. It allows a preferable resolution than classical logistic regression, which is unstable or not feasible at all in the situation of strong multicollinearity, small number of observations compared to the number of variables and missing values. In this paper, on the basis of hydrologic data from 1953 to 1998, PLS logistic regression is used to analyze the influence of monthly maximum five days precipitation and monthly maximum flow volume in YnnKt.se River on the disaster degree of flood in PoYang Lake. Furthermore, a model to predict the possibility of disaster degrees in PoYang Lake is established by use of this method. The research shows that PLS Logistic Regression has a fairly good applicability in the relevant fields.更多还原