I'm trying to use a while loop to delete the final row in an array every time it loops, however, it's not working

Basically, using the file TensileData.xlsx that I have attached to the question, I have to use a loop, work backwards through the entire data set reducing the number of points to find the equation of the elastic region. This can be done by calculating the r^2 value until the r^2 value is larger than 0.9997. However, when I try to create write the code for this I am unable to do it. I'm not completely sure where to begin, however, I have made a start. When I run the code I receive the error: Index in position 1 is invalid. Array indices must be positive integers or logical values.

Error in lab8t5 (line 26)

strain(numRows_strain, :) = [];

The code that my question relates to is part b onwards, however, I have put all the code here in case it is helpful.

clear all; close all; clc

% Part a.
data = importdata('TensileData.xlsx');
strain = data.data(:, 1);
stress = data.data(:, 2);

figure(1)
plot(strain, stress, 'r*')

% Part b.
numRows_strain = length(strain);
numRows_stress = size(stress, 1);
[a0, a1, r2] = linreg(strain, stress);
slope = a1;
constant = a0;
error = r2;

while error <= 0.9997
    [a0, a1, r2] = linreg(strain, stress);
    numRows_strain = numRows_strain - 1;
    numRows_stress = numRows_stress - 1;
    strain(numRows_strain, :) = [];
    stress(numRows_stress, :) = [];    
end

Answer 1

[Dyuman Joshi]

Firstly, use readmatrix or readtable instead of importdata.

Secondly, you can define variables from function calls, instead of reassinging them. And as mentioned before, do not use error as a variable name, because it is a built-in MATLAB function.

%Directly use this
[slope, constant, err] = linreg(strain, stress);

Thirdly, you are not updating the variable error in the while loop, so depending upon its value, the while loop will not run or run forever.

% Part a.
data = readmatrix('TensileData.xlsx');
strain = data(:, 1);
stress = data(:, 2);

figure(1)
plot(strain, stress, 'r*')

% Part b.
[slope, constant, err] = linreg(strain, stress);

while err <= 0.9997
    %Removing the last row
    strain(end, :) = [];
    stress(end, :) = [];  
    %Computing the r_square value again
    [~, ~, err] = linreg(strain, stress);
end

Since I do not have the linreg() function, I can not run the above code, but it should do what you asked for - delete the last row till the condition is met.

[Image Analyst]

You can make it more robust like this

while err <= 0.9997 && ~isempty(strain) && ~isempty(stress)
    % Remove the last row from each matrix:
    strain(end, :) = [];
    stress(end, :) = [];  
    
    % Bail out if either matrix is empty.
    if isempty(strain) || isempty(stress)
        break;
    end
    
    % Compute the r_square value again.
    [~, ~, err] = linreg(strain, stress);
end

[Dyuman Joshi]

Good point @Image Analyst

Answer 2

[Richard Burnside]

I think you could just do something vectorized like this:

idx = find(r2>0.9997);

stress(idx) = [];

strain(idx) = [];

Also I think "error" is a Matlab function so you should avoid using it as a variable.