You are part of the data science team at a multinational beverage company, tasked with developing a machine learning model to forecast the profitability of a new line of naturally flavored bottled waters in various locations. You are given historical data that includes product types, sales volumes, expenses, and profits across all regions. Considering this information, what should you use as inputs (features) and output for your model to accurately predict profitability?