Skip to main content
Sentiment filter

Sentiment explained, challenges, and filter location.

Updated over a week ago

Our sentiment analysis algorithm uses the latest in Machine Learning and AI, like Deep Learning and Pretrained Language Models (PLM), similar to those used by Google and Microsoft. This brings our AI as close as possible to truly understanding mentions.

Our model is language agnostic, meaning it can detect sentiment in nearly any language with high accuracy. And we analyze over 90 languages, including Chinese, Arabic, and Greek!

We trained our model on tens of thousands of examples to ensure top-notch quality. Our dedicated teams constantly work on improving its accuracy.

While detecting sentiment can be tricky due to sarcasm and context, our algorithm has improved from 61% to 95% accuracy.



We hope this article has been useful to you! If you have any further questions or need help, please don't hesitate to get in touch. You can reach us via email at help@brand24.com or send us a message using the chat box on the bottom-right side of your dashboard.

Did this answer your question?