Python：对组中心值 n SD 内的数字进行分组

Question

I have a list of multiple floats that looks something like this

mylist = [10, 10.2, 10.5, 11, 15, 15.3, 15.4, 16, 27, 27.4, 28, 28.1, 28.2]

I want to group the values that are close to each other. For eg. I want to group values from 10 to 11 into the average of the 4 values. I am having a hard time identifying the central values and then selecting values to left and right that would fall in the group. How could I do this?

Answer 1

How about this using defaultdict :

In [1]: from collections import defaultdict

In [2]: group = defaultdict(list)

In [3]: mylist = [10, 10.2, 10.5, 11, 15, 15.3, 15.4, 16, 27, 27.4, 28, 28.1, 28
   ...: .2]

In [4]: for val in mylist:
   ...:     group[int(val)].append(val)
   ...:     

In [5]: group
Out[5]: 
defaultdict(list,
            {10: [10, 10.2, 10.5],
             11: [11],
             15: [15, 15.3, 15.4],
             16: [16],
             27: [27, 27.4],
             28: [28, 28.1, 28.2]})

It does not need sorted input. Also, it preserves the order of related values

assuming, I correctly understand your requirement.

Answer 2

I sounds like you want a general method, probably something like:

from scipy.stats import binned_statistic

data = [10, 10.2, 10.5, 11, 15, 15.3, 15.4, 16, 27, 27.4, 28, 28.1, 28.2]
stats, edges, binarray = binned_statistic(data,data,bins=4)

edges    # Is the boundary values that split the data evenly into 4 bins. 
binarray # Shows which numbers in your original array belong to which equal sized bin. 
         # Note that nothing belongs to bin-3 because the gap is too wide.