Question

I’m trying to do analysis on datasets that have a varying number of columns, but I need to use 3 columns per use of my equation. I want to use all columns in my dataset (with the exception of the first and last as they will not work) But I need to select a center column, the column before, and the column after. I need to implement these equations:

lower_actual = lower_original - dark;
lower_avg = sum(lower_actual)/length(lower_actual);
lower_gain = lower_avg./lower_actual;

upper_actual = upper_original - dark;
upper_avg = sum(upper_actual)/length(upper_actual);
upper_gain = upper_avg./upper_actual;

middle_actual = middle_original - dark;

x1 = lower_actual;
x2 = middle_actual;
x3 = upper_actual;
y1 = lower_gain;
y3 = upper_gain;
y2 = (((x2-x1).*(y3-y1))./(x3-x1))+y1;
interpolate = y2.*middle_actual;

In these equations the variables correlate to:

lower = column before
middle = center column
upper = column after
dark = first column in data set

Answer 1

Suppose you work on a matrix M, e.g.

M = randn(10, 7);

Just loop over the columns skipping the first one and the last one.

interpolate = zeros(size(M, 1), size(M, 2) - 2);
dark = M(:, 1);
% Loop over the columns. Do not use the first and the last column as
% center columns.
for idx = 2 : size(M, 2) - 2
    lower_original = M(:, idx);
    middle_original = M(:, idx + 1);
    upper_original = M(:, idx + 2);

    % Your computations.
    lower_actual = lower_original - dark;
    lower_avg = mean(lower_actual);
    lower_gain = lower_avg./lower_actual;

    upper_actual = upper_original - dark;
    upper_avg = mean(upper_actual);
    upper_gain = upper_avg./upper_actual;

    middle_actual = middle_original - dark;

    x1 = lower_actual;
    x2 = middle_actual;
    x3 = upper_actual;
    y1 = lower_gain;
    y3 = upper_gain;
    y2 = (((x2-x1).*(y3-y1))./(x3-x1))+y1;

    interpolate(:, idx - 1) = y2.*middle_actual;
end

The maximum column index which is fetched is size(M, 2) - 3 + 2 == size(M, 2) - 1, i.e. the last column will be skipped.