ABSTRACT
The daunting complexity of biological networks and the multifactorial nature of common disorders pose a significant challenge for traditional biomedical approaches to understanding and diagnosing disease. The growing volume of healthcare and “omics” data suggests machine learning techniques, which are well-adapted to finding patterns in large and complex datasets, may offer a powerful tool for improved disease classification. This review shares a perspective on the unique obstacles this field confronts, the tools available to deal with them, key precautions for the application of machine learning in healthcare, and an assessment of the future outlook for this field.
