Per metric feedback
The API also offers detailed feedback on each of the Pronunciation, Fluency, Vocabulary, Grammar and Cohesion metrics. Such feedback can be used to explain why a user got a low score on a specific metric. Such feedback is available in the following nodes:
fluency
This node contains fluency metrics and sub-scores for the overall utterance and for each segment (sentence) within the utterance.
word_score_list[]
This node contains pronunciation scores and metrics for each word, syllable, and phoneme within the utterance.
grammar
This node contains grammar metrics, errors and feedback for the overall utterance.
vocabulary
This node contains vocabulary metrics, errors and feedback for the overall utterance.
coherence
This node contains coherence metrics, errors and feedback for the overall utterance.
Details regarding fluency and pronunciation scoring were explained in detail in earlier parts of this documentation. Each of the Grammar, Vocabulary, and Coherence feedback metrics has 3 sub-elements:
score: on a scale of 1 to 10
level: an interpretation of the score as low/mid/high
message: a feedback message if the score is low.
In the next few sections, we will review the grammar, vocabulary and coherence scores in detail:
Last updated