Salience 5.1.1 Feature Highlight: Composite Word Recognition

  22 s

As a result of our work with tokenizing Chinese, a language in which there are no word boundaries and one word may be comprised of two or more characters, we've developed the ability to pull apart multiword hashtags like #ilovelexalytics, or #iphonesucks. We are then able to use the resulting string to correctly calculate sentiment, making analyzing sentiment in tweets and content from other social media platforms that support hashtags easier and more useful. 

Categories: Announcements, Language