ABSTRACT
The convergence of biomedical Big Data, machine learning, and prediction research is defined as predictive analytics in medicine. This chapter reviews the principles of prediction research as the scientific framework of predictive analytics and the emergence of machine learning methods as the key tool set in this field. It also discusses how to build a prediction model with a step-by-step guide that goes from the formulation of the prediction problem all the way to the validation and implementation of the model using examples to illustrate the process. As defined by the Prognosis Research Strategy Group, the prediction research framework can be further subdivided into three areas of study: (a) prediction factor research, (b) prediction model research, and (c) stratified medicine. One key consideration when developing a prediction model is what data will be used to develop the model and what data will be used to validate and estimate its accuracy.
