Hi again,
Thanks for your input; I'm starting to use generators to some extent.
Say we have a series of numbers:
x = randn(100)
and values beyond some criteria should be considered as outliers, but
only where there's at most 3 (or some other integer) consecutive values
beyond the criteria. The outliers should then be replaced with linearly
interpolated values using the non-outliers.
outs = abs(x) > 0.5 # identify candidate outliers
# where(outs) would tell the location of the outliers
Here we could use something like:
from itertools import groupby
for val, g in groupby(outs):
if outs:
len(list(g)) # just getting sequence length here
to work with each outlier sequence, but then the indices in x are
forgotten. They'd be needed to get the interpolates. I feel there's
idioms that I should learn about to handle this situation more
effectively.
Cheers,
--
Seb
--
https://mail.python.org/mailman/listinfo/python-list
