This chapter shows how learning techniques can be applied to real-world data by predicting inpatient hospital readmissions and determining practical timeframes for predicting expenditures. It utilizes unsupervised time series clustering to distinguish high utilizers’ behaviors in Emergency Department and hospital inpatient visits, which is a method known as computational behavioral phenotyping. Other methods include a range of statistical and machine learning models, such as decision trees, random forest, support vector machines, and neural networks. To test the inpatient readmission predictive models, the chapter also utilizes the Healthcare Utilization Project US Nationwide Readmissions Database from the Agency for Healthcare Research and Quality. Regularized regression, also known as the least absolute shrinkage and selection operator, is usually the default approach in many biomedical supervised machine learning tasks. Deep learning has made great progress in machine learning tasks like object image classification, voice recognition, and natural language processing.