Implementation in Python of some of the statistical methods provided by
"asurv", the survival analysis software.
The original asurv software can be found at:
python-asurv can be found at: https://sourceforge.net/projects/python-asurv/
At the moment the only method implemented is the Schmitt binning method
(Schmitt, J. H. M. M. 1985; http://adsabs.harvard.edu/abs/1985ApJ...293..178S)
and probably, this method will be the only one implemented. If you are
interested in a regression method that can handle censored data in both axis
without the problems of the Schmitt method (arbitrary binning, statistical
properties not known, problems with small samples) you should have a look to
the Akritas-Thiel-Sen method. It is explained in the book "Nondetects And Data
Analysis: Statistics for Censored Environmental Data" (Wiley-Interscience,
2005, ISBN: 9780471671732). There is an implemetation of the method in R
(http://www.r-project.org) in a package called NADA
(http://cran.r-project.org/web/packages/NADA/index.html). The method can be
interfaced from Python using RPy (http://rpy.sourceforge.net/).
Python-asurv depends on numpy (http://numpy.scipy.org/).
To install python-asurv enter:
python setup.py install
If you use this software for your work you should cite one of the following
articles explaining the method used and the software (ASURV) in which this
software is based:
* Feigelson, E. D. and Nelson, P. I. "Statistical Methods for
Astronomical Data with Upper Limits: I. Univariate Distributions",
Astrophyscal Journal 293, 192-206, 1985.
* Isobe, T., Feigelson, E. D., and Nelson, P. I. "Statistical Methods
for Astronomical Data with Upper Limits: II. Correlation and Regression",
Astrophysical Journal, 306, 490-507, 1986.
* LaValley, M., Isobe, T. and Feigelson, E.D. "ASURV", Bulletin
Amercan Astronomical Society (Software Reports), 22, 917-918, 1990.