CAJ | 학술논문

模具行业客户数量相对较少但订单较多，针对用客户人口学数据进行客户流失分析的不足，基于订单信息建立模具业的订单流失预测模型。针对模具业订单样本分布极不平衡及其产生的不同错分代价的问题，提出一种基于Logistic回归的多元分类器方法，将此方法应用于某大型模具企业的订单流失分析，并与传统的Logistic回归预测算法进行比较，实验结果表明该方法能有效解决模具行业的订单流失分析问题。
모구행업객호수량상대교소단정단교다，침대용객호인구학수거진행객호류실분석적불족，기우정단신식건립모구업적정단류실예측모형。침대모구업정단양본분포겁불평형급기산생적불동착분대개적문제，제출일충기우Logistic회귀적다원분류기방법，장차방법응용우모대형모구기업적정단류실분석，병여전통적Logistic회귀예측산법진행비교，실험결과표명해방법능유효해결모구행업적정단류실분석문제。
The number of customers in the mold industry are limit but the orders are adequate, in response to the unavailability of customer demographics while doing customer churn analysis in the mold industry,order churn prediction model of mold industry based on customer contractual information and business data was set up. To deal with the challenge of a highly skewed class distribution between churn and non-churn and different classification cost it cause, a multi-classifier approach based on logistic regression was proposed. Then, the proposed method was applied to a mold enterprise, by comparing with the model build by traditional logistic regression, results suggest that the proposed method exhibits satisfactory predictive effectiveness in the mold industry.