李鹏鹏,周丹阳,姜朝明,喻湄霁,刘伟,王涛
LI Pengpeng,ZHOU Danyang,JIANG Chaoming,YU Meiji,LIU Wei,WANG Tao

• 1:国网浙江省电力有限公司台州供电公司
• 2:西华大学电气与电子信息学院

摘要(Abstract)：

为减少投诉风险发生,提出一种基于随机森林算法的95598工单投诉预测方法,实现对95598工单的直接投诉预测与转化投诉预测。首先,对95598历史工单进行数据预处理;其次,在充分考虑历史工单的供电地区、时间、天气、前期工单事因、重复来电和投诉倾向等情况的基础上,建立了基于随机森林算法的95598电力服务投诉工单预测模型。以某市全年95598工单数据为例,建立了该市的95598电力服务投诉工单预测模型,并以Weka 3.8数据挖掘软件为测试平台,对所建立的模型进行测试,并与其他数据挖掘算法的预测性能进行了对比分析。结果表明,该方法能够实现对95598投诉风险的有效预测,投诉预警效果良好。
To reduce the risk of complaints, the paper proposes a 95598 work order complaint prediction method based on Random Forest, which achieves its direct complaint prediction and transformed complaint prediction. Firstly, the data of 95598 historical work orders is preprocessed. Secondly, given the power supply area, time, weather, earlier work orders cause, repeated calls and complaints tendency of negative work orders, a negative work order prediction model of 95598 electric power service based on Random Forest algorithm is established. Finally, by use of the 95598 work order data of a city as an example, the negative work order prediction model of 95598 electric power service in the city is established. Taking the Weka 3.8 data mining software as the test platform, the model established in this paper is tested, and the prediction performance of other data mining algorithms is compared and analyzed. The results show that this method can effectively predict the risk of 95598 complaints and achieve favourable early warning against complaints.

关键词(KeyWords)： 数据挖掘;随机森林;投诉预测;电力服务;95598工单
data mining;Random Forest;complaint forecast;power service;95598 work orders

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 李鹏鹏,周丹阳,姜朝明,喻湄霁,刘伟,王涛
LI Pengpeng,ZHOU Danyang,JIANG Chaoming,YU Meiji,LIU Wei,WANG Tao

DOI: 10.19585/j.zjdl.202004010

参考文献(References)：

• [1]杨勇,严道波,徐敏,等.基于改进TFID F特征加权算法的95598投诉工单分类实现[J].电力与能源,2019,40(2):205-207.
• [2]罗欣,张爽.深度学习在电力潜在投诉识别分类中的应用[J].浙江电力,2017,36(10):83-86.
• [3]辛永,刘燕秋,黄文思,等.基于多模型融合的客户投诉风险预测方法[J].电力大数据,2018,21(11):31-37.
• [4]郁琛,吕友杰,段荣华,等.基于偏互信息法与支持向量机的覆冰闪络故障预警[J].电力系统自动化,2018,42(2):92-98.
• [5]信昆仑,刘龙,陶涛,等.基于用户水质投诉信息的供水管网污染源的追踪定位[J].天津大学学报(自然科学与工程技术版),2014,47(4):336-342.
• [6]吴潇雨,和敬涵,张沛,等.基于灰色投影改进随机森林算法的电力系统短期负荷预测[J].电力系统自动化,2015,39(12):50-55.
• [7]刘建荣.基于贝叶斯网络的公交停靠站服务质量改善策略研究[J].武汉理工大学学报(交通科学与工程版),2018,42(4):559-563.
• [8]黄衍,查伟雄.随机森林与支持向量机分类性能比较[J].软件,2012,33(6):107-110.
• [9]张华伟,王明文,甘丽新.基于随机森林的文本分类模型研究[J].山东大学学报(理学版),2006,41(3):5-9.
• [10]康有,陈元芳,顾圣华,等.基于随机森林的区域水资源可持续利用评价[J].水电能源科学,2014,32(3):34-38.
• [11]张颖,高倩倩.基于随机森林分类算法的巢湖水质评价[J].环境工程学报,2016,10(2):992-998.
• [12]李欣海.随机森林模型在分类与回归分析中的应用[J].应用昆虫学报,2013,50(4):1190-1197.
• [13]罗知林,陈挺,蔡皖东.一个基于随机森林的微博转发预测算法[J].计算机科学,2014,41(4):62-64.