Revealing the hidden knowledge in pathology big data: Machine learning to support clinical decision making for unknown infectious diseases and recognise biosecurity incursions in tropical Australia

Pathology departments generate massive data from the results of routine and targeted tests conducted to serve community health needs. With the availability of large pathology data collections, the application of sophisticated machine learning (ML) algorithms allows the detection of novel data patterns to characterise disease processes and monitor population health. We will apply recursive partitioning (trees and forests) and support vector machines (SVM) to large pathology data sets to: (a) investigate the biosecurity potential of linked community data, and; (b) assist early decision support for patients presenting with a tropical pyrexia of unknown origin (PUOs).