基于特征感知迭代的电网业务营销数据挖掘方法

摘要

The data mining of the "Internet plus electric" mode basically supports the precise power utilization.Current data mining methods for electricity enterprise marketing are rough description and low differentiation, which causes low compatibility of electricity marketing based on users behavior.Therefore, this paper proposes a data mining method based on iteration feature sensation for state grid enterprise marketing.Constructing the power consumer utilization model considering spatial feature, designing the mining model for marketing management decision-making tree, and filtering the redundant data feature to deduct the accurate marketing method for State Grid.The tests and simulations demonstrate that the proposed method has better performance of accuracy and lower data consuming.%"互联网+电力"的大数据挖掘能为精准用电提供基础支撑.现有电网营销数据挖掘对用户数据的挖掘粒度大,特征集区分度小,空间维度权值低,基于用户行为的电网营销策略准确度低.提出基于特征感知迭代的电网业务营销数据挖掘方法.构建结合空间特征的电力用户用电模型,设计营销管理决策树挖掘模型,对冗余数据特征进行过滤清洗,推导准确的电网营销行为.通过仿真和实验分析,验证新方法具有更好的预测精度和更低的数据消耗.