I think the soup_bowl is representing a general 3D view of the soup bow,unrelated to the x_train and y_train data in our lab.
If I am wrong, then correct me.
2nd question is:
These points in the contour plot and 3D plot of J(w,b) are indifferent, but when I click on those points (2nd and 3rd plots in the attached image) to see the x, y, and cost function values, then those don’t seem the same.
I think it happens because of user- limitation of following the same spot in the point of 3D graph. I attached the screenshot
Hey @farhana_hossain, look at the cost formula and see if it needs training data, will you do that?
No matter how one made the plot, it doesn’t change the formula. I mean I can make a plot with J = 2w^2 + 3b^2 + 4w + 5b +6 at my pleasure, but when we discuss cost in the context of machine learning, it needs training data.
If all you are interested is just how the plot was made, then check out its code implementation yourself. Okay?
I understand, but I think
**soup_bowl() is using the default values of x_train and y_train (already existing in prewritten code) .
And so, if I don’t put values of x_train and y_train, then the soup_bowl() is capable of producing the graph.
Now I see. Your focus was just the function implementation of soup_bowl, and not about the general cost function concept. I am sorry for misunderstanding you.
Perhaps you might still want to check out its implementation and see if, inside the function, it has read the same dataset of your concern or just used any default dataset.
