I was looking at Octave implementation of expm and fell over this line:
A(A == -Inf) = -realmax;
where A is a matrix.
What is the effect of the command?
I was looking at Octave implementation of expm and fell over this line:
A(A == -Inf) = -realmax;
where A is a matrix.
What is the effect of the command?
It puts the value -realmax in every entry of A which is -Inf.
For instance, let's say
A = [1 -Inf 3; Inf -Inf 4; 0 0 0];
Then the command above would change A to the following:
[1 -realmax 3; Inf -realmax 4; 0 0 0]