ガイド連動クイズ · 中 レベル

コールセンター音声分析における AI クイズ

引用された各研究の限界と範囲を含め、コールセンターの研究における音声認識、日記作成、感情分析をテストします。

関連するガイドパスコールセンターの音声分析における AI
質問 1 の 8

What output can automatic speech recognition provide for speech analytics?

すべての質問をプレビューする8 質問

クイズに答えると、答えと解説が表示されます。

  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?
    • 通話を別の言語に翻訳します
    • Segments the recording into speaker turns and assigns speaker labels
    • 通話の大きさを測定します
    • 録音を暗号化します
  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. 使用される言葉を超えて顧客の感情状態を検出するのに役立つ音響特徴はどれですか?
    • 録音のファイルサイズ
    • エージェントの従業員 ID
    • 電話キャリア
    • ピッチ、ペース、音量（韻律）
  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