Design and Implementation of Voice Intelligent Detection System for Power Grid Customer Service Based on Machine Learning
In order to improve the level of power intelligent customer service,a power customer service system based on artifi-cial intelligence is proposed.The system is divided into four parts:data layer,support layer,service layer and application lay-er.In order to effectively alleviate the time-consuming problem of manually combing voice information,a multi task integrated learning model is proposed to realize the emotion recognition of power customers,which can effectively save labor costs and au-tomatically extract valuable customer feedback information for power companies.An intention understanding model based on bi-directional recurrent neural network is proposed to further improve the availability and intelligence of intelligent customer serv-ice.In the experimental stage,the proposed model is verified by taking the power customer service call data provided by a pow-er company as an example.The results show that compared with the benchmark method,the performance of the proposed multi task integrated learning model is significantly improved,and the average accuracy in anger,happiness,neutrality and sadness is improved.In addition,the recognition accuracy of the proposed intention understanding model based on bidirectional recurrent neural network is 90.21％,the recall rate is 89.92％,and the F score is 90.07％.The simulation results further verify that the proposed model provides a reference for improving the quality of power service.
power systemintelligent customer serviceservice qualitydeep learning
万方数据
维普
