Hi,
In order to create a loop that processes multiple column vectors, you can encapsulate your existing code within a function and then iterate over each vector you want to process.
Firstly, you need to define the column vectors outside the code block.
% Define column vectors and store in a variable.
vectors = {ampboil5, amphand14, amphand5, ampice14, ampice5, ...
amproom14, amproom5, ampwarm14, ampwarm5, ...
roseboil5, rosehand14, rosehand5, roseice14, ...
roseice5, roseroom14, roseroom5, rosewarm14, rosewarm5};
Then, you can pass these vectors to a main function, for instance, calibrate_sensors as follows:
% Pass these vectors to calibrate_sensors() function to get the desired
% output
calibrate_sensors(vectors);
Lastly, you need to define the main and helper functions where you can substitute your code implementation.
% Define main and helper functions
function calibrate_sensors(vectors)
% Loop through each vector
for i = 1:length(vectors)
B = vectors{i};
B = remove_outliers(B);
calculate_uncertainties(B);
end
end
function B = remove_outliers(B)
N = length(B);
nu = N - 1;
Mx = mean(B);
Sx = std(B);
P = 0.95;
tnup = sqrt(nu / betaincinv(1 - P, nu / 2, 1 / 2) - nu);
xL = Mx - tnup * Sx;
xU = Mx + tnup * Sx;
% Remove outliers since <xL and >xU outliers
B = B(B >= xL & B <= xU);
end
function calculate_uncertainties(B)
N = length(B);
V = 1:N; % Example V, replace with actual values if different
Q = B; % Assuming Q is the same as B, replace if needed
al = (V * Q') / (V * V');
e = al * V - Q;
Sxy = sqrt(e * e' / (N - 2));
nu = N - 1;
P = 0.95;
tnup = sqrt(nu / betaincinv(1 - P, nu / 2, 1 / 2) - nu);
Dp = tnup * Sxy / sqrt(N);
% Zero error
Sa0 = Sxy * sqrt(V * V' / (N * (V * V') - sum(V)^2));
Dz = tnup * Sa0 / sqrt(N);
% Sensitivity error
Sal = Sxy * sqrt(N / (N * (V / V') - sum(V)^2));
ro = max(Q);
rx = ro / al;
Ds = tnup * Sal * rx / sqrt(N);
% Accuracy error
Da = sqrt((Dz^2) + (Ds^2));
% Total Calibration error
Dc = sqrt((Da^2) + (Dp^2));
end
I hope it will address your query.
Regards,
Zuber
