가이드 연계 퀴즈 · 중간 레벨

콜센터 음성 분석 퀴즈의 AI

인용된 각 연구의 한계와 범위를 포함하여 콜센터 연구에서 음성 인식, 분할 및 감정 분석을 테스트합니다.

관련 가이드 경로콜센터 음성 분석의 AI
질문 1 의 8

What output can automatic speech recognition provide for speech analytics?

모든 질문 미리보기8 질문

퀴즈를 풀면 정답과 설명이 나타납니다.

  1. What output can automatic speech recognition provide for speech analytics?
    • A machine-generated transcript for downstream text analysis
    • A guaranteed speaker identity
    • A final diagnosis of the caller’s emotional state
    • A payment authorization
  2. What does speaker diarization do in a call recording?
    • 통화를 다른 언어로 번역합니다.
    • Segments the recording into speaker turns and assigns speaker labels
    • 통화 소리의 크기를 측정합니다.
    • 녹음을 암호화합니다
  3. In IBM’s two-stream call-center emotion-recognition study, which information streams are combined?
    • Video and keystroke timing
    • Customer purchase history and account balance
    • Acoustic and lexical information
    • Only transcript punctuation
  4. 사용된 단어 이상으로 고객의 감정 상태를 감지하는 데 도움이 되는 음향 기능은 무엇입니까?
    • 녹음 파일 크기
    • 대리인의 직원 ID
    • 전화 통신사
    • 음조, 속도, 음량(운율)
  5. In a two-stream call-center emotion system, what do acoustic features contribute?
    • Evidence from vocal delivery, such as how speech sounds
    • The caller’s verified identity
    • The account’s payment history
    • A guaranteed diagnosis of the caller’s emotion
  6. What task did Vidrascu and Devillers study for customer-service phone calls?
    • Predicting credit scores from call recordings
    • Speaker segmentation and emotion-related analysis of conversational turns
    • Generating live agent sales recommendations
    • Automating legal decisions about customer accounts
  7. Why might a call-center sentiment system combine lexical and acoustic features?
    • Acoustic features replace the transcript in every task
    • Lexical features reveal the phone carrier
    • The words and the way they are spoken provide complementary signals
    • Combining them guarantees the caller’s true emotion is known
  8. What does word-error rate compare in speech recognition evaluation?
    • Speaker identity against a customer account
    • Emotion labels against call duration
    • Call volume against agent schedule
    • Substitutions, deletions and insertions against the words in a reference transcript