Skip to main content
All CollectionsFrequently Asked Questions
What is Sentiment and how is Sentiment calculated?
What is Sentiment and how is Sentiment calculated?

Everything you need to know about what Sentiment is and how Sentiment is calculated in Kapiche.

Ryan Stuart avatar
Written by Ryan Stuart
Updated over a week ago

What is Sentiment Analysis?

Sentiment Analysis is the process of using Natural Language Processing (NLP) to automatically determine the sentiment expressed in a piece of text. This could include customer feedback from survey responses, emails, social media posts, and more.

Because people often express emotions in their writing, Sentiment Analysis is a powerful tool for businesses to monitor customer feelings toward their brand and track changes in sentiment over time.

How Does Kapiche Calculate Sentiment?

Kapiche uses Plumeria—our proprietary sentiment analysis algorithm—to assess sentiment both at an overall text level and at a sentence level.

How Plumeria Works

Plumeria has been trained to recognize sentiment (positive, negative, mixed, and neutral) based on the words used and how they interact within a piece of text.

For example:

  • "I absolutely love Kapiche!" → The combination of "absolutely" and "love" signals strong positive sentiment.

  • "I absolutely don’t love Kapiche." → The addition of "don’t" flips the sentiment to negative.

  • "The staff were amazing! Unfortunately, my food was cold." → The sentiment is mixed, as it contains both positive and negative elements.

How is Plumeria Trained?

Plumeria has been trained using large datasets where sentiment has been manually classified by humans. Over time, our machine learning algorithms allow Plumeria to become smarter and more accurate with continued use.

Understanding Sentiment Export Data in Kapiche

When you export data from Kapiche, you will see three key columns related to sentiment:

Sentiment Label

Sentiment Confidence

Sentiment Polarity

positive

0.530840511

positive

neutral

0.873669733

neutral

neutral

0.773240018

neutral

mixed

0.712901951

negative

positive

0.61523122

positive

mixed

0.649870431

positive

positive

0.751314729

positive

negative

0.663238043

negative

What Each Column Represents:

  • Sentiment Label – The overall sentiment classification assigned to the text (Positive, Negative, Neutral, or Mixed).

  • Sentiment Confidence – The probability score (ranging from 0 to 1) that the assigned sentiment label is correct. Higher values indicate stronger confidence in the classification.

  • Sentiment Polarity – A new value that identifies the dominant sentiment when the overall sentiment is Mixed.

How Sentiment Polarity Works:

  • Sentiment Confidence is the renamed version of the old "Polarity" column. It represents how confident Plumeria is in its sentiment classification (on a 0-1 scale).

  • Sentiment Polarity is a new value that identifies the dominant sentiment when the overall sentiment is Mixed.

  • How it works:

    • When the sentiment is Positive, Negative, or Neutral, Polarity will always match the Sentiment Label.

    • When the sentiment is Mixed, Polarity identifies the stronger sentiment (Positive or Negative).

    • In the Mixed case, Polarity and Confidence together help you determine what the dominant sentiment is and how strongly it is expressed.

Questions? 🤔

If you have any questions about Sentiment analysis (or anything else!) you can get in touch with us any time by hitting the blue chat button to your right 👉

Did this answer your question?