耿俊成,张小斐,袁少光,万迪明.基于逻辑回归模型的电力客户停电敏感度评分卡研究与实现[J].电力需求侧管理,2018,20(3):46-50 |
基于逻辑回归模型的电力客户停电敏感度评分卡研究与实现 |
Research and implementation of users’outage sensitivity score card based on logistic regression model |
投稿时间:2017-10-17 修订日期:2017-12-28 |
DOI:10.3969/j.issn.1009-1831.2018.03.012 |
中文关键词: 逻辑回归模型 停电敏感度 评分卡 证据权重转换 信息值 |
英文关键词: logistic regression model power failure sensitiv ity score card weight of evidence information value |
基金项目:国家电网公司科技项目“基于营配信息贯通 |
|
摘要点击次数: 2463 |
全文下载次数: 8 |
中文摘要: |
对当前基于逻辑回归模型的客户敏感度分析模型输出结果复杂、不易理解与应用的问题,提出了一种基于逻辑回归模型的电力客户停电敏感度评分卡实施方法。从电网公司营销业务系统、95598系统提取客户停电敏感相关属性数据,通过信息值筛选出具有高预测力的属性;计算各属性的证据权重转换值,运用逻辑回归模型构建客户停电敏感度分析模型;基于停电敏感度分析模型的输出参数和各属性证据权重转换值,构建客户停电敏感度评分卡。以某电网公司实际数据对该方法进行了验证,取得了良好效果。 |
英文摘要: |
Aiming at the problem that the output of the users’sensitivity analysis model based on the logistic regression is complex and not easy to understand and apply, an implementation method of the users’outage sensitivity score card based on logistic regression model is proposed.First, it extracts the attribute data related with users’outage sensitivity from the State Grid marketing business sys tem and 95598 system, and selects the high predictive attributes through the information value.Then, the weight of evidence of each attribute value is calculated, and users’outage sensitivity analysis model is constructed Finally, the users’outage sensitivity score card is constructed based on the output parameter of sensitivity analysis model and the weight of evidence values.The method is verified by the actual data and the verification achieves good results. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |