A method for controlling a printer includes receiving a set of first variables for the printer, determining, using a machine-learning program, one or more second variables based on the set of first variables, and controlling a printer to perform a print job based on the set of first variables and the one or more second variables. The first and second variables may correspond to different types of printer settings. The machine-learning program may determine the second variables using a classifier model that may prevent partial curing or drying, print media deformation and/or other defects.