Commit 27743786 authored by Baptiste Bauvin's avatar Baptiste Bauvin
Browse files

Working

parent 824c7a0b
Pipeline #5284 passed with stages
in 1 minute and 21 seconds
WIP Multiview Generator
========================
.. |pipeline| image:: https://gitlab.lis-lab.fr/dev/multiview_generator/badges/master/pipeline.svg
:alt: Pipeline status
.. |license| image:: https://img.shields.io/badge/License-BSD%203--Clause-blue.svg
:target: https://opensource.org/licenses/BSD-3-Clause
:alt: License: New BSD
.. |coverage| image:: https://gitlab.lis-lab.fr/dev/multiview_generator/badges/master/coverage.svg
:target: http://dev.pages.lis-lab.fr/multiview_generator/coverage/index.html
:alt: Coverage
|pipeline| |license| |coverage|
Multiview Generator
===================
This package aims at generating customized mutliview datasets to facilitate the
development of new multiview algorithms and their testing on simulated data
representing specific tasks.
Understanding the concept
-------------------------
The main idea of the generator is to build several monoview sub-problems that
.. image:: _static/fig_rec.png
:width: 100%
:align: center
This package aims at generating mutliview datasets according to multipe paramterers :
an error matrix, that tracks the error of each view on each class, a redundancy float that controls the ration of examples that are well described by all the viewsw,
similarly the lutual error float controls the one that are misdescribed in eavery views.
Structure
---------
......
......@@ -9,7 +9,7 @@ n_features: 3
n_informative: 3
class_seps: 10
class_weights: [0.125, 0.125, 0.125,]# 0.125, 0.125, 0.125, 0.125, 0.125,]
mutual_error: [0.4, 0.4,0.4]
mutual_error: [0.2, 0.2,0.2]
redundancy: [0.5,0.4,0.4]
complementarity: [0.1, 0.05,0.05]
name: "demo"
......
......@@ -24,6 +24,11 @@ To install SMuDGE, clone the gitlab repository and run
include_tuto3
documentation
Read me
=========
.. include:: readme_link.rst
Indices and tables
==================
......
.. include:: ../../README.rst
\ No newline at end of file
__version__ = '1.0.dev0'
__version__ = "0.0.dev0"
......@@ -159,7 +159,7 @@ class MultiViewSubProblemsGenerator:
self.gen_report(save=False)
meta_data_grp.attrs["description"] = self.report
meta_data_grp.create_dataset("example_ids", data=np.array(
meta_data_grp.create_dataset("sample_ids", data=np.array(
self.example_ids).astype(
np.dtype("S100")), dtype=np.dtype("S100"))
......@@ -322,6 +322,7 @@ class MultiViewSubProblemsGenerator:
n_classes=self.n_classes,
shuffle=False,
random_state=self.rs,
cov=3.,
**self.sub_problem_configurations[
view_index])
......@@ -421,8 +422,6 @@ class MultiViewSubProblemsGenerator:
Randomly selects the examples that will participate to complementarity
(well described by a fraction of the views)
"""
n_comp = [int(self.complementarity_level[class_index]*self.n_views)
for class_index in range(self.n_classes)]
if ((self.complementarity * self.n_examples_per_class)[0] > np.array(
[len(inds) for inds in self.available_init_indices])).any():
raise ValueError(
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment