How can I identify the column with the greatest average in a numpy array?

Question

Given the following array:

    complete_matrix = numpy.array([
    [0, 1, 2, 4],
    [1, 0, 3, 5],
    [2, 3, 0, 6]])

I would like to identify the column with the highest average, excluding the diagonal zeros. So, in this case, I would be able to identify complete_matrix[:,3] as being the column with the highest average.

I [added this answer](http://stackoverflow.com/a/17420604/832621) the takes into account the exclusion of the diagonal of zeros — Saullo G. P. Castro, Jul 02 '13 at 08:23

score 2 · Accepted Answer · edited May 23 '17 at 12:11

Is this question different from the one here: Finding the row with the highest average in a numpy array

As far as I understand, the only difference is the matrix in this post isn't a square matrix. In case this was deliberate, you could try using weights. Since I do not understand your intent fully, the following solution assigns 0 weight to zero entries, 1 otherwise:

numpy.argmax(numpy.average(complete_matrix,axis=0, weights=complete_matrix!=0))

You can always create a weight matrix where the weight is 0 for diagonal entries, and 1 otherwise.

score 1 · Answer 2 · edited May 08 '14 at 20:01

1

Something like:

import numpy

complete_matrix = numpy.array([
    [0, 1, 2, 4],
    [1, 0, 3, 5],
    [2, 3, 0, 6]])

print complete_matrix[:,numpy.argmax(numpy.mean(complete_matrix, 0))]
# [4 5 6]

edited May 08 '14 at 20:01

Saullo G. P. Castro

56,802
26
179
234

answered Jul 01 '13 at 15:59

Jon Clements

138,671
33
247
280

How can I identify the column with the greatest average in a numpy array?

2 Answers2

Linked