Calculate mean of a matrix columnwise if a condition is met

I am stuck on a small coding issue.

I have a big matrix "A" and a scalar value "b" and I want to determine,

Expected value = Mean ( A -b) if A < b

for example if,

A = [10 10 10;10 5 5;10 5 5]
b = 10
C = b - A

C = [0 0 0; 0 -5 -5; 0 -5 -5]

Calculate mean = Expectation [C / (A<b ]

calculations of mean should be columnwise.

[Cris LaPierre]

Could you complete your example? What is the resulting value(s) of C for the example you shared?

[susman]

done!

Answer 1

[hosein Javan]

A = [10 10 10;10 5 5;10 5 5]
b = 10
C = b - A
C(A>=b) = nan % replace all data out of specified range by nan
mean(C,'omitnan') % mean columns omitting nan

[susman]

Thats absolutely perfect! Thanks

[hosein Javan]

you're welcome. don't mention it.