摘要
为了推动客户失效信息的校准与基于电力联系网络的稳定服务能力,提高电力主动通知的准确性和渠道服务的粘性,提出电力客户自助查询系统智能IVR语音自动播报设计方法。对采集的语音播报信号进行多分辨融合滤波处理,以多尺度的小波特征分解方法提取信号的关联频谱特征量,采用频谱特征分析方法,智能提纯和检测分析语音信号,根据语音信号的特征聚类结果,实现对电力客户自助查询系统智能IVR语音自动播报和优化特征辨识。仿真结果表明,采用该方法进行电力客户自助查询系统智能IVR语音播报的稳定性能较好,语音信号检测准确率较高,提高了电力客户自助查询智能服务水平。
In order to promote the calibration of customer failure information and the stable service ability based on power contact network to improve the accuracy of power active notification and the stickiness of channel service, a design method of intelligent IVR voice automatic broadcast for power customer self-service query system is proposed. The collected voice broadcast signal is processed by multi-resolution fusion filtering. The multi-scale wavelet feature decomposition method is used to extract the associated spectrum feature quantity of the signal, and the spectrum feature analysis method is used to intelligently purify, detect and analyze the voice signal. According to the feature clustering results of the voice signal, the intelligent IVR voice automatic broadcast and optimized feature identification of the power customer self-service query system are realized. The simulation results show that the stability of IVR voice broadcast is good, the accuracy of voice signal detection is high, and the intelligent service level of power customer self-service query is improved.
作者
杨欣
严军
蒋道乾
YANG Xin;YAN Jun;JIANG Dao-qian(Yunnan Power Grid Liability Co.,Ltd.,Kunming 650000,China)
出处
《信息技术》
2022年第1期114-118,125,共6页
Information Technology
关键词
电力客户
自助查询系统
智能IVR语音
自动播报
特征聚类
power customers
self-service inquiry system
intelligent ivr voice
automatic broadcast
feature clustering
