c2_w1_L-3_coffeeroasting normalization


can anyone please tell why the values are felled even to less than zero for temperature after normalization, I am not able to understand why such a huge difference with mean and variance is there, when they are into usage that made temperature, duration differ so much to negative values.

Hello @Suprem_Mohan,

It was performing the z-score normalization. To see the difference that it can bring, try to normalize these numbers [100, 200, 300] on a piece of paper.

For how to compute z-score normalization, check out the C1 W2 Video “Feature scaling part 2” starting around 2:50. Note that there is a subtraction operation.

Good luck!
Raymond