What languages are supported?

Types of support

Stratifyd supports various languages in a number of ways. 

• Stratifyd supports all basic text analytics functions natively in 27 languages including English, Chinese, Japanese, Spanish, Italian, German, and French. 

• We provide a sentiment dictionary for each of the 27 supported languages so that you can perform sentiment analysis in addition to all other NLU analyses.

• For other languages, so long as we can tokenize the data (i.e. there are spaces between words), you can perform all other NLU analyses. For languages with no spaces between words such as Chinese, Japanese, and Korean, Stratifyd must create a special tokenizer. 

• The pre-trained Neural Sentiment Model supports only English, and we provide a separate Chinese version. 

Sentiment dictionaries

Stratifyd includes a sentiment dictionary for each of the 27 supported languages. You can also use customize sentiment dictionaries to account for terms that are specific to your organization. Stratifyd supports all NLU, text analytics, and sentiment analysis functions natively in the following languages:

For non-supported languages, you can use the Translate Text feature to translate from over 100 languages into a supported language, and then run sentiment on that. 

1. Arabic

2. Chinese

3. Dutch

4. English

5. Filipino

6. French

7. German

8. Greek

9. Hausa

10. Hebrew

11. Hindi

12. Indonesian

13. Irish

14. Italian

15. Japanese

16. Korean

17. Persian

18. Polish

19. Portuguese

20. Romanian

21. Russian

22. Shona

23. Spanish

24. Swahili

25. Swedish

26. Turkish

27. Welsh

Key

This is the key for the language table below.

• Fully supported means that we support the language and have worked extensively to provide robust sentiment dictionaries.

• Supported means that we support the language and provide at least a rudimentary sentiment dictionary, but we are working on improvements. Please let us know of specific issues that we can improve and word lists to add to the sentiment dictionary.

• Blank means that, provided the language is tokenized (that is, there are spaces between the words), we assume that it would work for basic analyses, but there is no sentiment dictionary and it has not been tested.

• Microsoft NMT, one of two supported translation services, means that we support translation to and from the language if you select that translation service.

• Google Translate, one of two supported translation services, means that we support translation to and from the language if you select that translation service.

Any further questions on which languages we support? Don’t hesitate to reach out!