Question about the regularized term in the logistic regression
Dear TAs,
In the implementation of reg-regression model, we implemented the regularization term with a factor as\(\frac{1}{2} w^{2}\). However, in the new version description you published just before deadline we recognize that the factor should be \(w^{2}\). So we are wondering is our implementation still reasonable? These two terms can lead to the same answer with adjusting the parameter lambda.
From the project1_description.pdf, you should implement the function from the labs which according to Lab05 is \(\lambda w^2\). So your implementation for \(\frac{\lambda}{2} w^2\) may not pass the auto grading system. We will look at the code to see if anything fails the test, but I may not guarantee that it is always the case. If you lost code quality points for this reason, then you can send us email to explain.
Question about the regularized term in the logistic regression
Dear TAs,
In the implementation of reg-regression model, we implemented the regularization term with a factor as\(\frac{1}{2} w^{2}\). However, in the new version description you published just before deadline we recognize that the factor should be \(w^{2}\). So we are wondering is our implementation still reasonable? These two terms can lead to the same answer with adjusting the parameter lambda.
Thank you so much in advance!
From the project1_description.pdf, you should implement the function from the labs which according to Lab05 is \(\lambda w^2\). So your implementation for \(\frac{\lambda}{2} w^2\) may not pass the auto grading system. We will look at the code to see if anything fails the test, but I may not guarantee that it is always the case. If you lost code quality points for this reason, then you can send us email to explain.
2
Add comment