基于Matlab的语音情感特征值研究

Research on Customer Satisfaction Based on Speech and Emotional Eigenvalue

客服部门存有大量闲置的客服语音录音,这些录音占据一定规模的存储,不加以利用,只能荒废。通过Matlab程序对西安科技大学京东校园实训中心的客服服务语音录音进行了归类、整理、分析,并将其短时能量、语音节拍的特征值进行了量化处理,计算出与之相对应的最大值、最小值和均值,从而分析得出了其特征值与客户满意度之间的相关关系。该结论不仅为研究客户满意度与客服语音情感特征值之间的关系奠定了基础,也为网络在线购物平台提高客户满意度提供了一定的帮助,同时也有利于其经济效益的提升和实现优势性经营。

Customer service department has a large number of idle customer service recording voice which occupies a certain size of the storage, unless properly utilized, it can only be abandoned. The programming of Matlab is used to store and analyze the recording and phonetic voice of customer service of Jingdong campus training room of Xi′an University of Science and Technology. It also quantifies the characteristic value of the short-time energy and the tempo, and calculates their corresponding maximum, minimum and mean values, so as to analyze the relationship between the emotional eigenvalue and the customer satisfaction. This conclusion not only provides a basis for studying the relationship between customer satisfaction and customer service voice emotion value, but also gives some help to the online shopping platform to improve their customer satisfaction, meanwhile, it is also conducive to their economic efficiency and superiority operating.