How to make the estimation error and average it correctly?

Is the code correct for the whole part especially the estimation error and average error part?

Answer 1

[Paul Hoffrichter]

    for sd = 0:10:50
        PN=Po+(sd.*gaussnoise);
    end

In each iteration of this inner loop, PN is computed with a new sd value. But when this loop exits, the only value that PN has is the one that corresponds to sd = 50.

Is it your intent that each experiment have one random gaussnoise value, and then create errors for the 5 PN values corresponding to sd = [0 10 20 30 40 50]?

If so, then you have to move the end further down so that the different 5 different PN values are captured for your 5 tests per experiment.

[gp]

Yes, but even if i move to downwards it is still the same

[Paul Hoffrichter]

>> it is still the same

"it" does not compute.

I do not understand how you can have different values of sd which results in different values of PN and still have the same results. You should post your code and print out the results to show the problem.

Answer 2

[KALYAN ACHARJYA]

It seems the code considering the mean error for the last experiment iteration. You may same the E of all experiment and later get the error. Here E as array.

E=zeros(1,10); %preallocation
for experiment = 1:10
    .......
     E(experiment)=(errM).^2;
     %....^
end
Em=mean(E);
disp(['Em=' num2str(Em)]);