Help Center

Sentiment Analysis in CX Social

Last Updated: Jul 13, 2017 11:33AM EDT
Sentiment Analysis in CX Social
For every mention that lands in CX Social we automatically determine the sentiment. We label each mention with a positive, negative, neutral or unknown sentiment. In this support article we give you a full explanation on how sentiment analysis works.

ŸŸ●  How does Sentiment Analysis work?
ŸŸ●  How many languages do we support?
ŸŸ●  Editing Sentiment for mentions manually
ŸŸ●  Sentiment Insights in CX Social
How does Sentiment Analysis work?
For CX Social we developed an automatic sentiment detection algorithm which determines the sentiment polarity class (negative, neutral, positive) of each text mention in CX Social.
We use state of the art classification algorithms. Specifically we use entropy classification of texts represented as “batch of words” enriched with the source and the length of texts. Basically this means that we look at the different classifications (=batch) in a text and ignore the syntactic information such as word order. Next the algorithm computes the probability per batch and determines the polarity class. Our algorithm is trained on a combination of manually annotated and semi-automatically mined training examples.
Sentiment Analysis in CX Social has greatly improved over the last couple of years and performs at least as good as other services that are publicly available. We achieve an accuracy and precision of 75% which on average is almost as good as humans. 
How many languages do we support?
In CX Social we support the following languages for automatic sentiment analysis: English, Dutch,  French, German, Greek, Italian, Spanish, Portuguese, Turkish and Arabic.
Editing Sentiment for mentions manually
You always have the option to overwrite the automatic sentiment that we assigned to a mention in CX Social. You can easily do this by selecting the mention in the Inbox and clicking on the smiley face in the mention engagement bar. Here you will be able to select a positive, neutral or negative sentiment.  

Sentiment Insights in CX Social
CX Social also always you to drill down on mentions  based on sentiment. If you want to find out how many mentions were positive or negative over a certain period of time you can see this in the Insights.
You can either use the filter field and filter on mentions via the Sentiment filter that is available or you can use the pre-made widget in the Insights and get a quick summary of the sentiment in your topic. If you want to read and learn more about sentiment Insights in CX Social you can read this here.

Contact Us
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
Invalid characters found