If you notice that the system does not properly detect the language of incoming emails or chat messages, use the following tips to resolve the problem.
The following causes might prevent the system from detecting the language properly:
- The expected language might not be enabled for natural language processing (NLP) in Prediction Studio settings.
- The analyzed text might too short. For example, when you analyze texts in similar languages, a small sample might not provide sufficient data for the NLP engine to detect the language correctly.
- Depending on the number of languages that you use, perform one of the
following actions to get the expected output:
Options Actions Using a single language
In the text analyzer settings, specify the language that you want to use.
The default setting in the text analyzer is to detect the language automatically. If you are using this text analyzer to detect a single language, and the language is not detected, you can configure the text analyzer to detect a specific language.
For more information, see Configuring language detection preferences.
Using multiple languages
- Check whether the language that you are expecting is enabled for NLP in the Prediction Studio settings.
- Check whether the analyzed text is long
enough.Language detection might not work properly on a short piece of text. The text should be at least three sentences long. For example, if you are using an email channel, the email body should contain a lengthy text of at least three sentences for the language to be detected properly.