06. Summary

Recap

In this lesson, we introduced the course by providing two examples of how wearables aim to be more than just fitness devices. We started with an anecdote of when a wearable device alerted someone to a serious medical decision. And we followed that up with a discussion about the largest clinical study ever conducted with wearables, the AHS.

Wearable data science spans a few broad domains:

• signal processing.
• machine learning.
• electronics.
• sensors.
• some clinical domain expertise.

This course assumes that you have some basic knowledge of machine learning. (However, if you don’t, you can always catch up by spending more time with the Further Resources sections at the end of each lesson!)

In this course, we begin with introductions to signal processing and sensors to provide you with some of the necessary background knowledge to be successful in this field. From there, we dive into a few case studies in biomedical time series algorithms that process wearable data. Each particular case study will begin with any of the clinical or physiological domain knowledge necessary to complete the exercises.

Glossary

• Inclusion Criteria: Characteristics that potential study subjects must have for them to be included in the study.
• Exclusion Criteria: Characteristics that disqualify potential study subjects from participating in a clinical study. e.g., Some common ones are being under 18 or pregnant.
• Classification Accuracy: A metric for evaluating the performance of a classifier -- the fraction of classifications that are correct. For rare events (like atrial fibrillation), this metric is unsuitable. For example, a classifier that classifies every data point as healthy would have a classification accuracy of 99%, as around 1 percent of the population has atrial fibrillation, but would be relatively useless.
• Precision: The fraction of positive classifications that are correct.
• Recall: The fraction of positive elements that are classified correctly as positive.
• Primary Endpoint: The metric being used to answer the question that the study seeks to ask. For drug trials, this would be markers of the disease the drug seeks to treat. For example, the primary endpoint for a study on the effectiveness of a new statin in preventing heart attacks would be the number of heart attacks in the test group compared to a control group. The number and types of participants enrolled in a study are designed with the primary endpoint in mind.