Cheng-i Wang
Cheng-i Wang
Hi @urinieto and @bmcfee, I was comparing codes between my own implementation of scluster and the one in the repo. And I find out that the current version in the...
It would be this part here in the original implementation, [laplacian_segmentation](https://github.com/bmcfee/laplacian_segmentation/blob/94a2c347d841554cf4d1a3b0f39b666f23ecf576/code/segmenter.py#L492-L514), where an entropy measurement is used to estimate the best "k". I can put together a merged version between...
Hey @urinieto, without deep investigation into my own codes, the first point seems fine. About the second one, I agree with you I should comply to the option of pre-defined...
That makes sense. I will adjust it if you have not.
The `suffix_method` is unique to `MO` and not applicable to either `VMO` or `FO`. For `FO`, the two options for `suffix_method` will be the same due to symbolic comparison. Since...
I think it was setup like that so when casting something like `y = numpy.array(oracle.symbol)` you do not have to worry about the type, but at the end I think...
Looks good just by a glance. Can you do a pull request?
The [query function ](https://github.com/wangsix/vmo/blob/master/vmo/analysis/analysis.py#L245) was used in that publication. I will add an example notebook using that function. In the mean time, please experiment with the function and let me...
My original intent was to have the basic FactorOracle as the parent class. FO and MO inherited FactOracle. For VMO, it`s a variant of MO that is still under experiment....
What do you mean by modifying them to be uniform? Thanks!