Ever have a regression model where the coefficients don't make sense? I've been trying to predict electricity and gas consumption from daily activity schedules but a simple linear regression kept saying that demands should go down the more an activity is performed. Fortunately I found the nnls package and show here how you can use it to...