3. OUR SOLUTIONThe solution we developed to predict risk for lead poison-ing is based on a variety of data sources. We obtained data from the Chicago Department of Public Health that consistsof blood lead level (BLL) tests and home-inspection records,combined that with housing records and other public data(described in detail below), and built a classifier to predict the risk of lead poisoning. The city of Chicago has adoptedthe CDC definition of lead poisoning of a BLL of 5 g/dL.Our system consists of the following components:1. Data Integration and Cleaning2. Feature Generation3. Model Selection and Training4. Model Validation5. Deployment and ImplementationThe next several sections describe each of the componentsin more detail.