Recursive function with for loop inside?

Hi Comunity.

I come asking for your wisdom

Im writing a recursive function that has a for loop inside.

The idea is to find a precise factor that must be inside the range [5, 10]

This factor is used to calculate the number of points inside a matrix that are above a threshold.

To achieve this I'm iterating over the range untill the number of elements over threshold is 1

For example, lets asume that the factor im looking for is 9.945

Iterating over 5 to 10 is not precise, so im calling the function recursively sending as arguments more precise limits, in this case 9 and 10

The function is working well: the correct factor is returned in the "recursive breaking condition", but due to the loop, continues in the first calls of the recursion, the result is overwriten.

Some ideas of how to achieve this correctly?

A simple idea would be to iterate from 5 to 10 whit a little step, like 0.00001, but the idea is to make it efficient. If the factor is inside the range [9, 10], previous values must be skipped quicly

Here's the code

temp is the matrix that contains numeric values.

Threshold is calculated with as 30 * factor

fact = factor(10, 5, 1, temp, 30);

function f = factor(upLim, downLim, step, temp, threshold)
    for i = downLim : step: upLim
        numNR = numel(temp(temp > threshold * i))
        if step == 0.0001 % Return condition
            f = i
            return
        elseif numNR == 1
            f = factor(i, i-(step), step/10, temp, threshold); % Here the range is restricted and the step reduced
        end
    end
end

When finding the factor value = 9.945 with the step 0.0001, the function is returned, but in the previous calls the for loop continue iterating.

How to totally break the loop when intering the return condition?

Thx a lot for the ideas

[Mathieu NOE]

hi

use break

[Felipe Bayona]

How to use break in the previous calls of the function?

[Jan]

What are typical inputs for temp? Can you explain the purpose of the function in clear words?

I assume the calculations should be much easier, e.g. f = max(temp) / threshold. Or maybe the 2nd largest element matters.

[Felipe Bayona]

Hi Jan

Temp is a 3D matrix with values from 1 to 1000. The idea is to find the threshold where there are only 1 value above.

For various statistical reassons, the threshold is one base number times a factor

B x Factor. In thios particular case B = 30, so the threshold must be a multiple of 30, where factor is between 5 and 10

[Jan]

I assume Mathieu would insert the break after the "f = factor(..." line. Wouldn't this solve the problem?

But I still think that you should be able to solve this algebraic using the value of the threshold and the one or two largest elements of the array.

Answer 1

[David Hill]

I think you just need a break after your recurssive call.

function f = factor(upLim, downLim, step, temp, threshold)
    for i = downLim : step: upLim
        numNR = numel(temp(temp > threshold * i))
        if step == 0.0001 % Return condition
            f = i
            return
        elseif numNR == 1
            f = factor(i, i-(step), step/10, temp, threshold); % Here the range is restricted and the step reduced
            break;%I think you just need a break here
        end
    end
end

[Felipe Bayona]

Thx David.

I was not sure how and instruction after the recursive call would be executed.

Thank you very much