一种基于Logistic回归的银行保险消费行为影响因素分析与预测

Analysis and Forecast Model of Bancassurance Consuming Behavior Based on Logistic Regression Method

运用Logistic回归方法分析了银行保险消费行为的影响因素,并构建了预测模型,然后利用保留样本进行了实证检验.通过对样本数据过离散、空单元、多元共线性、特异值、特殊影响案例的诊断,以及对模型的拟合程度和预测效果的检验,验证了所构建的模型具有较好的拟合程度和预测效果.

Logistic regression method is applied to analyze effect factors of bancassurance consuming behavior and build a forecast model. Then the model is tested with reserved sample. Through the diagnosis of overdispersion, zero cell count, multicollinearity, outliers, influential observations and test of level of fitness and predictive accuracy, the model is proved to be adequate and effective,indicating that 12 factors are of significance in statistics science.