Disease Prediction Using Deep Learning Methods

Diagnostic error is an increasing threat to public health that contributes to a higher mortality rate of patients than any other preventive hospital adverse event. In a clinical setting, diagnostic error is responsible for 6 to 17 percent of adverse events. Previous works have shown that an individual physician’s diagnosis accuracy was 62.5% and increases to 85.6% if teams of physicians worked together in diagnosing the patient. Worse can be said for complex cases and infrequent cases, where the diagnostic phase is lengthy and ridden with errors. The aim of this project was to develop a machine learning model that can support doctors to accurately predict infrequent diseases as well as common diseases. We designed a multimodal one-shot learning method tha integrates both note text and numerical lab test data to diagnose a patient. We use the late fusion strategy that is commonly adopted in multi-modal learning and siamese network for one-shot learning.

This project was part of the 2470 Deep Learning course.