Table 16 RNN and CTC speech transcription accuracy.

From: A zero-shot LLM framework for multimodal grievance classification, urgency scoring, and abuse detection in civic feedback systems

Condition

WER (%)

Confidence

Quiet, Short (<5s)

4.8

0.94

Quiet, Long (>10s)

6.1

0.91

Noisy (SNR >15dB)

9.5

0.87

Noisy (SNR <10dB)

13.2

0.78

Mobile Mic Input

7.9

0.89

Back to article page