How does the sentiment analysis work?
Prodsight uses machine learning to perform sentiment analysis on your customer feedback to help you understand how your users feel about various aspects of your product or service.
Prodsight automatically analyzes each mention in your workspace for sentiment, categorizes it and applies one of three labels:
- Positive - mention text was mostly positive
- Neutral - mention text was mostly neutral
- Negative - mention text was mostly negative
The sentiment classification is based on the overall opinion, feeling, or attitude expressed in a mention.
Based on the sentiment associated with individual mentions we calculate various metrics which allow you to understand sentiment at a topic-level, project-level, subtopic-level and overall.
Net Sentiment is a single metric that helps you quickly gauge the sentiment associated with a given concept.
Net Sentiment is calculated by subtracting the percentage of negative mentions from the percentage of positive mentions.
Net Sentiment = % Positive Mentions - % Negative Mentions
Net Sentiment ranges from -100% to 100%. Anything above 0% is considered positive an anything below - negative.
Sentiment Breakdown is a set of metrics that describe how sentiment is distributed across a topic or a project.
- Counts - a simple tally of the positive, neutral and negative mentions associated with a topic/project/subtopic.
- Percentages - a percentage expression of positive, neutral and negative mentions associated with a topic/project/subtopic.