Share via

Low Confidence level of Language Identification

Amper, Charwin (Contractor) 65 Reputation points
2024-04-30T04:02:25.82+00:00

Hi,

I was testing the this file , which is in English language, and somehow the language identification returned with Low confidence level for en-US locale. I used both continuous and recognize once option.

Are there options I can set, to always ensure High confidence level for the right locale?

Azure AI Speech
Azure AI Speech
An Azure service that integrates speech processing into apps and services.
1,437 questions
Azure AI Language
Azure AI Language
An Azure service that provides natural language capabilities including sentiment analysis, entity extraction, and automated question answering.
364 questions

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Umar 160 Reputation points
    2024-04-30T05:31:10.9433333+00:00

    To ensure a high confidence level for language identification, you can try providing more context or longer text samples for analysis. Additionally, check if the file contains any unusual formatting or mixed languages that might affect the accuracy of the language identification process. You can also explore using custom language models or fine-tuning existing models to better suit your specific needs.

    0 comments
  2. santoshkc 5,000 Reputation points Microsoft Vendor
    2024-04-30T06:07:00.6233333+00:00

    Hi @Amper, Charwin (Contractor),

    Thank you for reaching out to Microsoft Q&A forum!

    To improve the confidence level of language identification, you can increase the length of the text, use high-quality text which means free of spelling errors, grammatical errors, and other issues that could confuse the language identification service. Also, you can try creating a custom language model using the Custom Speech Service to improve the accuracy of language identification.

    I hope you understand! Thank you.

    0 comments