search

lamtung16 / ML_ChangepointDetection

0 stars 0 forks source link

squared hinge versus square loss for training #5

Closed tdhock closed 1 month ago

tdhock commented 9 months ago

hi @lamtung16 here is a result figure showing that linear model trained by minimizing square loss (lm.2) is about the same as linear model trained by minimizing squared hinge loss (linear.2), can you please create a similar figure using your code, for debugging purposes? Untitled

lamtung16 commented 9 months ago

Hi Dr. @tdhock, this is my figure, I used different shapes for my results and paper's results, just for differentiation. I have exactly the same results in BIC and constant, slightly different in linear model. image

tdhock commented 9 months ago

that is good progress, thanks for sharing. now can you please add all of your other variants to this same figure?

lamtung16 commented 9 months ago

I added them into the figure image