Dcoumentation

[Francesco Tuveri]

I'm trying to understand the following functions that are included in the bob.learn.em package: linear_scoring, t_norm and z_norm. Is it possible to have a bit more documentation about them? In particular I don't understand what's the purpose of linear_scoring and how to use it. For what concerns t_norm and z_norm I know what they should do, but I don't understand what kind of inputs are expected. Thanks.

[Manuel Günther]

Dear Francesco,

although I am not an expert on this, I think the ``linear_scoring`` is a fast way to compute the (an approximation of the) log-likelihood ratio between several model files (GMMs) and several probe observations (GMMStats) using the given universal background model (UBM). What the other two parameters in http://pythonhosted.org/bob.learn.em/py_api.html#bob.learn.em.linear_scoring are, I don't know. I guess they have proper default values.

I agree that the documentation of the function is very poor. @tiago: Would it be possible for you to improve the documentation? I have assigned an according issue to you :-D

The other two functions do not really belong to bob.learn.em (as they do not make use of the EM algorithm) and are inside this package for historical reasons. They are used for ZT score normalization.

Cheers

Manuel

[Amir Mohammadi]

Hi Francesco,

The documentation of bob.learn.em is improved:
https://www.idiap.ch/software/bob/docs/latest/bob/bob.learn.em/master/index.html

Maybe the new documentation answers your question.

Best,
Amir