Merge pull request #52 from glouppe/joss

[MRG] Submission for JOSS

carl/__init__.py

@@ -56,7 +56,7 @@
 import sklearn.base
 from sklearn.base import clone as sk_clone
 
-__version__ = "0.0"
+__version__ = "0.2"
 
 
 def _clone(estimator, safe=True, original=False):

paper.bib

@@ -0,0 +1,8 @@
+@article{Cranmer:2015-llr,
+      author         = "Cranmer, Kyle and Pavez, Juan and Louppe, Gilles",
+      title          = "{Approximating Likelihood Ratios with Calibrated
+                        Discriminative  Classifiers}",
+      year           = "2015",
+      eprint         = "1506.02169",
+      archivePrefix  = "arXiv",
+}

paper.md

@@ -0,0 +1,35 @@
+---
+title: 'carl: a likelihood-free inference toolbox'
+tags:
+  - likehood-free inference
+  - density ratio estimation
+  - Python
+authors:
+ - name: Gilles Louppe
+   orcid: 0000-0002-2082-3106
+   affiliation: New York University
+ - name: Kyle Cranmer
+   orcid: 0000-0002-5769-7094
+   affiliation: New York University
+ - name: Juan Pavez
+   orcid: 0000-0002-7205-0053
+   affiliation: Federico Santa María University
+date: 4 May 2016
+bibliography: paper.bib
+---
+
+# Summary
+
+Carl is a likelihood-free inference toolbox for Python. Its goal is
+to provide tools to support inference in the likelihood-free setup,
+including density ratio estimation algorithms, parameterized supervised
+learning and calibration procedures.
+
+Methodological details regarding likelihood-free inference with calibrated
+classifiers can be found in the companion paper [@Cranmer:2015-llr].
+
+Future development aims at providing further density ratio estimation
+algorithms, along with alternative algorithms for the likelihood-free setup,
+such as ABC.
+
+# References