Biochemical and Molecular Basis of Pediatric Disease

Abstract

Background Clinical laboratories generate a large number of test results, creating opportunities for improved data management and the use of analytics. Aggregate analyses of these data have potential diagnostic value but require labs to utilize computational tools like machine learning for the analysis of high-dimensional data, which contain a large number of variables (columns) for each observation (row). Machine learning can be used to aid decision-making, whether for clinical or operational purposes, using a variety of algorithms to analyze complex data sets and make reliable predictions. This chapter discusses key concepts related to big data, whose attributes require new technologies and analysis methods, and their application to pediatric laboratory medicine. Machine learning workflows, concepts, common algorithms, and related infrastructure requirements are also covered.