基于改进决策树的电力混合大数据实时采集方法

Real time power hybrid big data acquisition method based on improved decision tree

在电力混合大数据采集过程中,由于数据处理模式的影响,导致数据采集的量化误差较高。因此,提出基于改进决策树的电力混合大数据实时采集方法。利用智能映射技术,构建统一的混合数据采集模型。通过Sqoop脚本、Kafka,以及ftp三种方法的有机结合,完成电力采集混合数据实时交换。针对多级电力冗余数据,基于改进决策树算法,建立符合采集需求的数据处理模式。依托于采集的电力数据值设计自适应采集策略,实现电力混合大数据实时采集。实验结果表明:基于改进决策树算法的数据采集方法,对比两种方法,将量化误差降低了16%与22%,有效提升了数据采集质量。

In the process of power hybrid big data acquisition,the quantification error of data acquisition is high due to the influ-ence of data processing mode.Therefore,a real-time acquisition method for power hybrid big data based on an improved decision tree is proposed.Intelligent mapping technology is used to construct a unified hybrid data acquisition model.Real-time exchange of power acquisition data was completed through the organic combination of Sqoop script,Kafka,and ftp.For multi-level power redundancy data,a data processing mode meets the acquisition requirements based on the improved decision tree algorithm.Relying on the adap-tive collection strategy of collected power data values,we realize real-time collection of electric power mixed big data.The experimen-tal results show that,based on the data acquisition method of improving the decision tree algorithm,the quantification error is reduced by 16%and 22%,effectively improving the quality of data acquisition.