In the benchmark notebook, the formula for calculating the target variable is:

final_phase = final_phase.assign(target=np.maximum(final_phase.return_flow, 0) * final_phase.return_turbidity)

My understanding of the np.maximum function is that x2 is 0 so that we only return positive values. My question then is about return_flow. What exactly does it mean for return_flow to be negative? At first I thought that measured when liquid left the cleaning area, but not every Process_ID has a negative value in return_flow. I don’t know if this question is legitimate, but any pointers would be appreciated.

Thanks