[SciPy-user] gaussian_kde broken?

Flavio Coelho fccoelho at gmail.com
Tue Oct 3 13:46:40 CDT 2006


Hi,

I noticed a new bug in the combination Numpy1.0rc1/Scipy0.5.1:

#generate a sample from a normal distribution:
a=normal(0,1,10000)
from scipy.stats.kde import gaussian_kde as kde
d=kde(a)
d.evaluate(arange(-1,1,0.1))
---------------------------------------------------------------------------
exceptions.TypeError                                 Traceback (most recent
call last)

/home/fccoelho/Documents/Papers/YFgame/<console>

/usr/lib/python2.4/site-packages/scipy/stats/kde.py in evaluate(self,
points)
    116                 diff = self.dataset - points[:,i,newaxis]
    117                 tdiff = dot(self.inv_cov, diff)
--> 118                 energy = sum(diff*tdiff,axis=0)/2.0
    119                 result[i] = sum(exp(-energy),axis=0)
    120

TypeError: sum() takes no keyword arguments

Are there any fixes available for this?

thanks

-- 
Flávio Codeço Coelho
registered Linux user # 386432
---------------------------
"Laws are like sausages. It's better not to see them being made."
Otto von Bismar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.scipy.org/pipermail/scipy-user/attachments/20061003/0fc6728f/attachment.html 


More information about the SciPy-user mailing list