Closed GoogleCodeExporter closed 8 years ago
The help section of the numpy.var function says:
---------------------------------
Help on function var in module numpy.core.fromnumeric:
var(a, axis=None, dtype=None, out=None)
var(sample, axis=None, dtype=None)
Return the variance, a measure of the spread of a distribution.
The variance is the average of the squared deviations from the mean,
i.e. var = mean((x - x.mean())**2).
See also: std
---------------------------------------
So it looks like 'ddof' is not a valid argument for this - do I have the wrong
version of numpy?
Original comment by nickschu...@gmail.com
on 6 Jan 2010 at 9:37
Hi Nick,
You might have the wrong version of NumPy. The ddof parameter is the number of
degrees of freedom. It's been available with every version of NumPy I've used
over
the past two years. Red Hat Enterprise is very slow with updating NumPy and may
still
distribute 1.0.3. What version of linux are you using?
Try the following within ipython
import numpy
print numpy.version.version
When I try it on my machine, I get version 1.2.1.
Thanks,
Damian Eads
Original comment by damian.e...@gmail.com
on 8 Jan 2010 at 2:18
As I said in my original comment, I have version 1.0.1(!) - so it looks like
thats
the problem! I'll install the latest numpy and see if that fixes things...
Original comment by nickschu...@gmail.com
on 8 Jan 2010 at 9:16
Sorry, my fault, I did read your original posting carefully enough. ;) CentOS
probably distributes a very old version. Compiling NumPy should not be too
painful.
Compiling SciPy on the other hand will be nontrivial.
Original comment by damian.e...@gmail.com
on 8 Jan 2010 at 7:27
New version of numpy seems to have solved the problem. :D
Original comment by nickschu...@gmail.com
on 14 Jan 2010 at 10:26
Excellent!
Original comment by damian.e...@gmail.com
on 16 Jan 2010 at 3:12
Original issue reported on code.google.com by
nickschu...@gmail.com
on 5 Jan 2010 at 5:21