# End-To-End Prediction of Emotion From Heartbeat Data Collected by a Consumer Fitness Tracker

## TL;DR

This paper introduces a Bayesian deep learning model that predicts emotional valence from heartbeat data collected by consumer fitness trackers, demonstrating practical emotion detection in real-world settings with a new dataset and a peak F1 score of 0.7.

## Contribution

It presents the first end-to-end emotion classification model using PPG data from affordable fitness trackers, incorporating uncertainty estimation for real-world applications.

## Key findings

- Peak F1 score of 0.7 on the new dataset
- Demonstrates feasibility of emotion detection 'in the wild'
- Uses only heartbeat data from consumer devices

## Abstract

Automatic detection of emotion has the potential to revolutionize mental health and wellbeing. Recent work has been successful in predicting affect from unimodal electrocardiogram (ECG) data. However, to be immediately relevant for real-world applications, physiology-based emotion detection must make use of ubiquitous photoplethysmogram (PPG) data collected by affordable consumer fitness trackers. Additionally, applications of emotion detection in healthcare settings will require some measure of uncertainty over model predictions. We present here a Bayesian deep learning model for end-to-end classification of emotional valence, using only the unimodal heartbeat time series collected by a consumer fitness tracker (Garmin V\'ivosmart 3). We collected a new dataset for this task, and report a peak F1 score of 0.7. This demonstrates a practical relevance of physiology-based emotion detection `in the wild' today.