search

angleto / liblinear

BSD 3-Clause "New" or "Revised" License
0 stars 0 forks source link

coefficients larger than 1 #14

Open simsong opened 6 years ago

simsong commented 6 years ago

From @naarkhoo on May 13, 2016 9:25

I wonder, if coefficients can be larger than 1 even on the normalized data ? I am bench making liblinear with glmnet, and I see one of the coefficients is larger than one.

Thanks.

Copied from original issue: cjlin1/liblinear#22

simsong commented 6 years ago

From @cjlin1 on May 26, 2016 22:9

Do you mean the weight vector? If so, yes they can be larger than 1 naarkhoo writes:

I wonder, if coefficients can be larger than 1 even on the normalized data ? I am bench making liblinear with glmnet, and I see one of the coefficients is larger than one.

Thanks.

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub*

simsong commented 6 years ago

From @naarkhoo on June 10, 2016 22:44

Thanks for your answer. I have another issue, the glmnet solution is sparser than LIBLINEAR solution - I wonder if there could be any explanation ? :-) - thanks.

simsong commented 6 years ago

From @cjlin1 on August 3, 2016 22:45

If the same loss and the same parameters are used, I think they should give the same sparsity naarkhoo writes:

Thanks for your answer. I have another issue, the glmnet solution is sparser than LIBLINEAR solution - I wonder if there could be any explanation ? :-) - thanks.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*