Resonata's Sentiment calculation is based off of the contents of the posts that are analyzed by the Resonata system. The Resonata system will associate positive, neutral, or negative sentiment values to various words and phrases, which is used to calculate the overall sentiment of your community, various posts, threads, and authors.
The engine has been trained to capture the sentiment generically and it can be improved further to better match your community's sentiment. Here's how you train the sentiment engine:
- Highlight a word or a phrase that you think is positive or negative.
- Once the word/phrase is highlighted you'll be prompted with a dialog to choose the appropriate sentiment.
- Choose the appropriate sentiment.
- Repeat for other words/phrases in the post.
If you change a previously highlighted word/phrase it overrides the general training and the sentiment is updated for that post.
Although we've started highlighting the words/phrases that have a specific sentiment, the post itself still has an overall sentiment. Our sentiment engine does much more than just basic counting of words. Therefore you might see posts where there are positive and negative words highlighted while the overall sentiment of the post is neutral.
If you don't agree with the overall estimate of sentiment for a post, you can change the sentiment label in the "Labels" field. This helps train the sentiment engine and updates the sentiment counts for your search.