基于大数据的电力市场人工智能客服支持平台设计

Design of artificial intelligence customer service support platform in power market based on big data

针对当前95598客服服务系统智能化的需求,以及提高电力市场服务水平的要求,基于现有的95598平台,提出一种基于L S T M的95598话务工单预测算法,以此为95598客服提供技术支撑。为实现对95598人工智能客服的支持,在分析L S T M原理及结构的基础上,对L S T M的结构进行设计,包含整体预测的流程和L S T M神经网络的结构、激活函数等进行设计。最后,以浙江国网某供电局的话务工单数据为基础,对提出一种大数据的95598人工智能客服支持平台。为实现该平台,结合大数据技术思想,以营销系统、95598等话务工单系统数据作为来源,采用大数据中丰富的计算优势和算法对数据进行分析,最后挖掘的数据进行展示。而在算法实现部分,引入来电号码识别算法,以及结合95598数据对工单量进行预测,并给出测试结果。实验表明,结合浙江国网某供电局的数据,在3186条来电中,能准确识别3074条,准确率高达80.6%。同时借助L S T M算法,能准确预测话务需求。

In view of the current demand for the intelligence of 95598 customer service system and the demand for improving the service level of the power market,a big data 95598 artificial intelligence customer service support platform is proposed.In order to realize the platform,combined with the idea of big data technology,the data of marketing system,95598 and other traffic work order system are used as the source,and the rich computing advantages and algorithms in big data is used to analyze the data,and finally mining the data for display.In the part of algorithm implementation,the algorithm of caller ID recognition is introduced,and the work order quantity is predicted with 95598 data,and the test results are given.The experiment shows that 3074 calls can be accurately identified out of 3186 calls,with an accuracy of 80.6%.At the same time,with the help of LSTM algorithm,the traffic demand can be predicted accurately.