Anti symmetric adjacency matrix using sklearn

[adhikarirsr]

I have a code as follows:

xt=np.array([[0.753217, 
  0.0439285], [-0.827553, -0.244174], [0.0875135, -0.0413367], 
[-0.509302, 0.519792], [-0.0819656, 
  0.769458], [0.167709, -0.472054], [0.83912, -0.15233], [0.974581, 
  0.175885], [0.392318, 0.503732], [-0.196902, 0.265483]]);
adj = kneighbors_graph(xt, 2,metric='euclidean', mode='connectivity')
adj11 = adj.toarray()
(adj11==adj11.T).all()
=> False

I want to make an undirected graph. However ,sklearn gives me an antisymmetric matrix. Is there a way to get a symmetric adjacency matrix for the case above?

[rusty1s]

k-nearest neighbor graphs are always directed. You can make it undirected as follows:

adj11 = adj11 + adj11.T
adj11[adj > 1] = 1.

[adhikarirsr]

Thanks @rusty1s

That's what I was thinking too.
There's a typo in the second line. It should be: adj11[adj11 > 1] = 1 , right?

[rusty1s]

Yes :)