Reputation:
I have a set of numbers, just to make it easier:
import numpy as np
A = np.array([0, 1, 2, 3, 4, 5, ...])
I want to:
((0+1)/2 + (2+1)/2)/2 , ((1+2)/2 + (3+2)/2)/2, ....
I am not sure how to go about this. What I got so far:
B = (A[1:]+A[:-1])/2.0
I would expect to get:
B = [1, 2, 3, 4, 5 ...]
Any help would be great. Thank you.
Upvotes: 0
Views: 52
Reputation: 2726
>> import numpy as np
>> def pair_avg(a):
>> return (a[1:]+a[:-1])/2.0
>>
>> a = np.arange(10)
>> pair_avg(pair_avg(a))
array([1., 2., 3., 4., 5., 6., 7., 8.])
pair_avg does a running average of every pair in a. As I understand you want to do this twice. I think you should see this operation as a recursive application of the same function. If you always want to recurse twice, you can do
>> def oper(a):
>> b = (a[1:]+a[:-1])/2.0
>> return (b[1:]+b[:-1])/2.0
>>
>> oper(a)
array([1., 2., 3., 4., 5., 6., 7., 8.])
Of course, this is equivalent to some weighted sum of the original array elements. Every 2nd application of pair_avg
will just clip off two more end point values of the original array when you apply it to integers.
Upvotes: 1