Analysis Tools for Data Exercise

Somebody can help me with this exercise? Attached is a file with 10.000 ids to use as a learning population, divided into two classes: positive and negative. The data includes 50 demographic variables and 50 behavioral variables. I have to develop a model that will allow me to better identify the positive class that has a rare frequency. The file also contains 4.000 cases to be used as an evaluation population with the same set of variables, but for which the Y factor to be forecast is not revealed. The exercise is select 800 cases, seeking to maximize the selection of positive cases.