3 ms·
"In our linear regression problem, there was only one minimum. This made our error surface convex." Shouldn't it the the other way around?
by shitgoose 12y ago
"In our linear regression problem, there was only one minimum. This made our error surface convex."
Shouldn't it the the other way around?
- jmalicki 12y agoCorrect... if there is only one minimum, it is unimodal, but not necessarily convex. For instance, if there are inflection points, it will not be convex.