Sentiment Analysis in Activity Exports Follow

Sentiment provides rapid and accurate context as a summary of long-form text responses. The sentiment is powered by our integration with IBM Watson. This provides us with some unique metrics.

• Entities: These are people, places, events, and things mentioned in member comments
• Keywords: These are notable keywords (or phrases) identified in member comments

Each tab will display the entities or keywords associated with a particular comment and the corresponding sentiment scores.

Sentiment scores are measured from a 1 to -1 range, representing positive, neutral, and negative.

The Relevance Score is used to show how relevant a keyword or entity is to the entire comment. The higher the score, the more relevant or important the keyword or entity is in the comment as determined by the machine learning algorithms.