The Voice Research Company

Healthcare High Class of 2016 Machine Learning Superlatives

By Hannah Cloe

Mon Apr 03 2017

In 2016 inVibe rolled out a first of its kind machine learning integration that identifies vocal acoustic markers in a person’s speech to measure their emotional state.

The algorithm was trained to decode human vocal intonations into their underlying emotions in real-time using over 2.5 million emotion-tagged voice samples. The integration enabled inVibe’s voice-powered research platform to move beyond stated intentions or forced ranking to measure a concept’s effectiveness, to understand the emotional impact a condition has on a patient, or even predict future behaviors. We have already seen our clients use this technology in some pretty amazing ways and this article will serve as a sneak peek into what you can do with the results.

Emotions are labeled through a combination of 3 individual measures, as expressed by the respondent’s voice during an analyzed voice section.

  • Temper — Speaker’s overarching mood range. Temper ranges from gloomy or depressed to confrontational or aggressive.
  • Valence — Speaker’s overall sentiment. Valence ranges from negative to positive.
  • Arousal — Speaker’s energy level, activation, and stimulation. Arousal ranges from tranquil or bored to alert or excited.

Since launching this new capability, we have collected and analyzed over 13,000 segments of audio. From this dataset, we decided to do our version of a high school yearbook superlatives. Each of the following are based on the likelihood of a given respondent to display a certain emotion by condition for patients and specialty for physicians.

Most Likely to Demonstrate…

Arrogance & Supremacy:

Payers. I guess the rumors are true.

Cynicism:

Pharmacogenetics Specialists.

Passion & Creativity:

Parents of children with cerebral palsy.

Leadership and Charisma:

Parents of children with cerebral palsy. If that’s not inspiring, I do not know what is.

Defensiveness and Anxiety:

This was a near-tie between payers and women with dyspareunia.

Friendliness and Warmth:

The cardiology community owns this one. Cardiologists followed by patients with high cholesterol with less than a fraction of a percent behind.

Loneliness and Unfulfillment:

Parents of children with Hemophilia A.

Sadness and Sorrow:

Gastroenterologists at American College of Gastroenterology. (Must have been a riveting conference!)

Self-control and Practicality:

Genetic Counselors.

Other interesting tidbits…

Ad Testing:

Patients were far more likely to reply with hostility and anger during message testing than during exploratory research.

Mama Bear:

The highest score for arousal (excited) and temper (confrontational) were measured in the same participant at 99/100 for each, AND was paired with a valence (negative) score of just 6/100. This correlates to the most intensely hostile participant we have surveyed.

For us, this exercise unveiled a wealth of inspiration and excitement about the tool we have added to each and every study inVibe conducts, and hopefully reading it gets some thoughts flowing for you as well.

If you would like more information on this unique offering, would like to know how our clients have used it to make better business and marketing decisions, or would like to start a project, contact inVibe.

Thanks for reading!

Be sure to subscribe to stay up to date on the latest news & research coming from the experts at inVibe Labs.

Recently Published

  • How GPT-3 Unlocks Deeper Listening at inVibe

    inVibe's breakthrough NLP algorithm accelerates healthcare insights

    By Jeremy Franz

    Thu Jul 01 2021

  • /voice

    1. Why Voice
    2. Sociolinguistic Analysis
    3. Speech Emotion Recognition
    4. Actionable Insights
    5. Whitepapers
    6. The Patient Voice

    @social

    2021 inVibe - All Rights Reserved