Don't get which line of code compute Sigmoid's gradient

Course Q&A Machine Learning Specialization Supervised ML: Regression and Classification
1

Hey, should you need to manually get the derivatives of the sigmoid function?

Like in this one, didn’t get which line of code you figure our sigmoid’s derivatives

 for i in range(m):
        f_wb_i = sigmoid(np.dot(X[i],w) + b)          #(n,)(n,)=scalar
        err_i  = f_wb_i  - y[i]                       #scalar
        for j in range(n):
            dj_dw[j] = dj_dw[j] + err_i * X[i,j]      #scalar
        dj_db = dj_db + err_i
    dj_dw = dj_dw/m                                   #(n,)
    dj_db = dj_db/m                                   #scalar
2

The lectures give the equation for the partial derivatives. You only need to implement them.

3

It’s the line that does essentially ((f - y) * x). For the weights, it’s equation (3) here.

image
image1047×333 21 KB

4

(2) and (3) are essentially the same equation, if you grant that for the bias, (2) is equivalent to (3) if you set a feature value of 1.