Hi,
I am learning unit2 of “machine learning of robotics”. I think there is a typo:
the gradient of “b” should be:
-(2/n) * sum(y_i - (m*x_i + b))
please correct me if I am wrong.
the gradient of “b” should be:
-(2/n) * sum(y_i - (m*x_i + b))
sorry, I forget to offer the location of this formular. It is in unit2 - " 2.2 Linear regression - gradient descent"
This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.
Thank you Jialei for your indication.
The thing is that there is no typo in the text.
The one we have is:
And the one that you mean is:
As you can see, both of them are equivalent. On the first one, the minus is within the sum. Since that minus is equivalent to (-1) multiplying all the elements of the sum, this means that the (-1) can be taken out of the sum, then be put multiplying the 2. So the result is what you mention.
Those are two ways of describing the same. Pick the one that is clearer for you!