Q.2 (Logistic regression) simple calculation
For binary classification y = {0,1}, we obtain the following weights (0,0₁,0₂,03) after the training of
Logistic Regression, where each sample has three features (x₁,x₂,x).
0₁=10, 0₂=-10, 0₃=15.
Then, by using the sigmoid function, prediction of y for a given sample that has x₁ =1, x₂ =0, x₃ = 2.
Specifically,
fy=1 when
y=0 when
(z) > 0.5
(z) <0.5'
1
1+exp(-z)'
z= 0% + 0₁x₁ +0₂x₂ +0₃x₃;
What is the predicted value of y for the given sample when there is no regularization?
4
Following Q.2-1, what is the predicted value of y for the given sample when an extremely large value is
applied to the regularization term (e.g, λ=10^6)?