[Scipy-tickets] [SciPy] #1288: stats streamlining obrientransform

SciPy Trac scipy-tickets@scipy....
Sun Sep 26 08:25:21 CDT 2010


#1288: stats streamlining obrientransform
-------------------------+--------------------------------------------------
 Reporter:  josefpktd    |       Owner:  somebody
     Type:  enhancement  |      Status:  new     
 Priority:  normal       |   Milestone:  0.9.0   
Component:  scipy.stats  |     Version:  0.7.0   
 Keywords:               |  
-------------------------+--------------------------------------------------

Comment(by josefpktd):

 doing the example from the reference

 obrientransform returns wrong shape (1, nobs) instead of 1d (nobs,)


 {{{
 >>> counte = [0, 1, 2, 4, 9, 16]
 >>> countc = [5, 11, 9, 3, 2, 2]
 >>> num = np.arange(5, 11)
 >>> xe = np.repeat(num, counte)
 >>> xc = np.repeat(num, countc)
 >>> stats.f_oneway(xc, xe)  #Fstat should be 60.96
 (60.961525704809283, 8.5308207499060442e-011)
 >>> stats.f_oneway(stats.obrientransform(xc).ravel(),
 stats.obrientransform(xe).ravel()) #should be 1.29 0.2595
 (1.2935552889926629, 0.25977024023067896)
 }}}

 conclusion cannot reject hypothesis of homogeneous variances

-- 
Ticket URL: <http://projects.scipy.org/scipy/ticket/1288#comment:1>
SciPy <http://www.scipy.org>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list