首页>
外国专利>
Feature vector profile generation for interviews
Feature vector profile generation for interviews
展开▼
机译:用于访谈的传染媒介配置文件
展开▼
页面导航
摘要
著录项
相似文献
摘要
Aspects of the present disclosure concern cognitive interview management. Audio data from an interviewer in an interview is received. The audio data is processed by converting the audio data into structured text. The structured text is then analyzed to identify questions in the structured text. The structured text data is also analyzed to determine at least one tonal dimension, at least one class dimension, and at least one difficulty dimension of the structured text. A feature vector is then generated having the determined dimensions. A reference feature vector is received having at least one reference tonal dimension, at least one reference class dimension, and at least one reference difficulty dimension. The generated feature vector and reference feature vector are then compared. A determination is made whether the comparison satisfies a match threshold, and in response to the comparison satisfying the match threshold, the interviewer is selected for a panel.
展开▼