Abstract
Machine learning brings the hope of finding new biomarkers extracted from cohorts with rich biomedical measurements. A good biomarker is one that gives reliable detection of the corresponding condition. However, biomarkers are often extracted from a cohort that differs from the target population. Such a mismatch, known as a dataset shift, can undermine the application of the biomarker to new individuals. Dataset shifts are frequent in biomedical research, e.g., because of recruitment biases. When a dataset shift occurs, standard machine-learning techniques do not suffice to extract and validate biomarkers. This article provides an overview of when and how dataset shifts break machine-learning–extracted biomarkers, as well as detection and correction strategies.
Highlights
Computer aided diagnostic of thoracic diseases from X-ray dataset shift images has been shown to be unreliable for individbreaks learned biomarkers uals of a given sex if built from a cohort over-representing the other sex [Larrazabal et al, 2020]
Biomarkers are measurements that provide information machine-learning systems may fail on data from different about a medical condition or physiological state [Strimbu imaging devices, hospitals, populations with a different age and Tavel, 2010]
A primer on machine learning for shift – the empirical risk is a poor estimate of the expected error, and f will not perform well on individuals from the biomarkers target population
Summary
Computer aided diagnostic of thoracic diseases from X-ray dataset shift images has been shown to be unreliable for individbreaks learned biomarkers uals of a given sex if built from a cohort over-representing the other sex [Larrazabal et al, 2020]. Biomarkers are measurements that provide information machine-learning systems may fail on data from different about a medical condition or physiological state [Strimbu imaging devices, hospitals, populations with a different age and Tavel, 2010]. If the training examples are not representative of the target population – if there is a dataset. A primer on machine learning for shift – the empirical risk is a poor estimate of the expected error, and f will not perform well on individuals from the biomarkers target population
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
