文本挖掘技术在金融机构客户服务中的应用  

The Application of Text Mining Technology in Customer Service of Financial Institutions

作　　者：王宇[1] WANG Yu(China UnionPay Co.Ltd.,Shanghai 201201,China)

机构地区：[1]中国银联股份有限公司,上海201201

出　　处：《信息与电脑》2021年第9期175-180,共6页Information & Computer

摘　　要：目前,各大金融机构在日常运营过程中,会产生大量非结构化的语音或文本数据,在这些数据中往往蕴含了用户对企业产品、营销活动等最真实的反馈信息.然而,运营部门对这些数据的分析、挖掘主要依赖人工处理,工作量大,时效性不高,基于文本挖掘技术的智能化分析与挖掘能力,不仅会显著降低运营部门的日常运营成本,还能大大提升公司营销活动、产品设计的运营水平.笔者首先分析了企业运营过程中产生的文本表达特征,提出基于用户数据否定窗口的用户关键意图抽取方法,对运营文本数据进行预处理,然后采用聚类算法对数据进行归类,并提出基于关键词连接矩阵的聚类合并算法,对聚类结果进行二次合并,最后提出基于关键词评分的聚类摘要自动提取模型,抽取类簇的描述信息.最后,使用62、双12等重大营销活动期间的网络客服对话数据对模型进行验证,取得良好的效果.At present,the major financial institutions in the daily operation process,will produce a large number of unstructured voice or text data,which often contains the most authentic feedback information of users on enterprise products,marketing activities and so on.However,the analysis and mining of these data mainly rely on manual processing,with heavy workload and low timeliness.The intelligent analysis and mining ability based on text mining technology will not only significantly reduce the daily operation cost of the operation Department,but also greatly improve the operation level of the company's marketing activities and product design.This paper first analyzes the text expression features produced in the process of enterprise operation,and proposes a user key intention extraction method based on the user data negative window.The operation text data is preprocessed,and then the data is classified by clustering algorithm.In order to solve the problem of difficult cluster setting,we propose a clustering merging algorithm based on keyword connection matrix Finally,an automatic extraction model based on keyword score is proposed to extract the description information of clusters.Finally,we use the data of online customer service dialogue during 62,double 12 and other major marketing activities to verify the model and achieve good results.

关 键 词：文本挖掘 特征抽取 文本聚类 关键词连接矩阵 关键词评分 

分 类 号：G434[文化科学—教育学]