C1_W2_Lab02 optimal values of w don't make sense?

In C1_W2_lab02, 𝐰 and 𝑏 has been loaded with some initial selected values that are supposed to be near the optimal, but the 3rd value which is associated with the #floors is -ve…!! Does it make sense, #floors increase, the base value (b-value) decrease by a scale of -53.36?!!

Hello @mhemeda, the dataset has only 3 samples, and 3 samples are just too small to represent the real world, and consequently, whatever optimal model you trained from these samples can’t be too seriously compared with our common sense. I think the purpose of the lab is to show that now we are able to model with more than one feature.

1 Like

Hello Raymond,

Thank you so much… it’s clear now!


You are welcome, Mohammed.