Machine learning and data analytics

Abstract This chapter presents the current advancements toward the analysis of medical data to address the unmet needs for several diseases including patient stratification, detection of biomarkers, and effective treatment monitoring, among others. A complete workflow for machine learning and data analytics is presented starting from data preprocessing to the application of machine learning and data mining for the development of patient stratification models and treatment monitoring strategies, as well as the detection of prominent biomarkers. Emphasis will be given on the mathematical basis of both supervised and unsupervised machine learning algorithms, including regression analysis, Naive Bayes, artificial neural networks, and tree ensembles, among others. Furthermore, emphasis will be given on strategies to predict disease outcomes across distributed medical databases through incremental learning. Methods for visualizing complex data structures will be also presented along with popular machine learning frameworks and software tools including cloud-based solutions.