V-Blaze and V-Cloud Online Help (May 2020)

Confidence Scores

V‑Blaze provides a "confidence score" between 0 and 1 for the text that is produced. Confidence score is a measure of the system’s confidence that the output was transcribed correctly. This provides a very effective mechanism for filtering output. Confidence scores are available at the word, utterance, and call level.

Confidence is calculated for each word, then averaged up to the utterance level, then up to the call level. Confidence scores tend to decrease as calls become more noisy, and as the overall speech signal moves further away from the speech used to train the Voci acoustic and language models. Noise, compression artifacts, and accents all contribute to lower confidence scores.

Confidence Scores Comparisons Between Models

The confidence score is not an accuracy measure. It is a measure of how confident the ASR engine is that it has selected the word most likely to be correct out of all the words it believes a region of speech could represent. The confidence score is only relevant to the model being scored and should not be used to measure performance across multiple models. To measure performance across models, perform accuracy comparisons (described in Measuring Accuracy) against a reference set of data for each of the models.