Session 2

The dimensions of dw in multiple Linear Regression is:

• (1,1)

• (1,n)

• (1,m)

• (n,1)

if we use very Large lamda

 $\lambda$  in regulariztion ,we will face:

• Bias case

• Just right case

• Variance case

• other case 

In Variance case , the cost function of training is high & the cost function of testing is low

• True

• False

In which one of the figures do you think hypothesis overfits trainingset

• 
• 
• 
• 

which code is correct, check all answers, and choose correct

• sc_x = StandardScaler()

  x_train = sc_x.fit_transform(x_train)

  x_test = sc_x.transform(x_test)

• J = (1/2*m)*np.sum(y-y_hat)^2

• n = 5

  w= np.zeros( [ n, 1] )

• dw = (1/m) * np.dot(x,(y-y_hat).T)