Restructure ClassificationGAM 2
#272
Conversation
| ## A 2-element numeric vector specifying the prior probabilities for each | ||
| ## class. The order of the elements in @qcode{Prior} corresponds to the | ||
| ## order of the classes in @qcode{ClassNames}. This property is read-only. | ||
| ## A numeric vector specifying the prior probabilities for each class. The |
There was a problem hiding this comment.
No need to change the property docstring. It is always a 2-element numeric vector.
There was a problem hiding this comment.
I investigated matlab's working with Prior, so modified for MATLAB compatibility.
| switch (s.subs) | ||
| case 'Cost' | ||
| this.Cost = setCost (this, val); | ||
| this.Cost = val; |
There was a problem hiding this comment.
Keep my previous change utilizing a priveate method so it can also be called from subsasgn
There was a problem hiding this comment.
I tried using that, but when I tested, it was causing errors, I tried 2-3 ways but it didnt work, so came up with this.
There was a problem hiding this comment.
Well, it doesn't work because my previous code was wrong. Instead of
this.Cost = setCost (this, val);
it should have been
this = setCost (this, val);
There was a problem hiding this comment.
So should I keep the current changes, since it is working well right?
| Formula = []; | ||
| Interactions = []; | ||
| ClassNames = []; | ||
| Prior = "empirical"; |
There was a problem hiding this comment.
Prior must be a 2-element numeric vector. Don't initialize it like this
There was a problem hiding this comment.
Same mentioned reason of MATLAB compatibility.
| RSS = sum (res .^ 2); | ||
| endfunction | ||
|
|
||
| function this = setCost (this, Cost, gnY = []) |
There was a problem hiding this comment.
We need to keep this private method. See my comment above
There was a problem hiding this comment.
Same above mentioned reason for removing it.
| %! assert (a.DoF, [7, 7, 7]) | ||
| %! assert (a.BaseModel.Intercept, 0.4055, 1e-1) | ||
|
|
||
| ## Test Prior calculation |
There was a problem hiding this comment.
You also need to document this name-value argument in the constructor's help docstring
There was a problem hiding this comment.
## @item @qcode{'Prior'} @tab @tab A numeric vector specifying the prior
## probabilities for each class. The order of the elements in @qcode{Prior}
## corresponds to the order of the classes in @qcode{ClassNames}.
## Alternatively, you can specify @qcode{"empirical"} to use the empirical
## class probabilities or @qcode{"uniform"} to assume equal class
## probabilities.
Is this what you are talking about? If yes, then it is already added in the file.
2 participants
No description provided.