TOPICS (Click to Navigate)

Pages

Thursday, July 21, 2022

Machine Learning MCQ - Cost function of logistic regression is convex

Multiple choices questions in Machine learning. Interview questions on machine learning, quiz questions for data scientist answers explained, what is the cost function of logistic regression? Why the cost function used in linear regression cannot be used in logistic regression? Why the cost function of logistic regression is convex?

Machine Learning MCQ - Cost function of logistic regression is convex

< Previous                      

Next >

 

1. Which of the following statements about logistic regression are correct?

a) Logistic regression uses the squared error as the loss function

b) Logistic regression assumes that each class’s points are generated from a Gaussian distribution

c) The cost function of logistic regression is concave

d) The cost function of logistic regression is convex

Answer: (d) The cost function of logistic regression is convex


Gradient descent will converge into global minimum only if the cost function is convex in the case of logistic regression.

 

Can we use the same cost function used in linear regression for logistic regression?

If we use the cost function of linear regression (Mean Squared Error) in logistic regression, we end up with a non-convex function with many local minimums. In this case, it is very difficult to find the global minimum. This strange outcome is due to the fact that in logistic regression we have the sigmoid function around, which is non-linear (i.e. not a line).

   

< Previous                      

Next >

 

************************

Related links:

Why the cost function of logistic regression is convex?

Why can't we use the cost function of linear regression in the case of logistic regression?

Can we use the same cost function of linear regression in logistic regression?

What will happen if you use linear regression's cost function in logistic regression?

Machine learning solved mcq, machine learning solved mcq

No comments:

Post a Comment