How to apply formula with if statement

Hello Community,

First time poster so please go easy on me.

I am also very new to matlab. I have just started doing some data fitting and want to do some work using first principles before using the in built data fitting tools.

I wish to fit my data to a log-normal distribution using schmeisers approximation for phi^(-1) and am able to get this done using excel rows to do the mathematics for me. I would rather use a MATLAB if statement to do the mathematics to automate this process in the future and I have had an attempt at this as you can see below

if freqvector < 0.0028
    phiminus1 = -(0.2+((1-freqvector).^(0.14)-(1-(1-freqvector)).^0.09)/0.1596)
elseif freqvector > 0.9972
        phiminus1 = 0.2 + ((freqvector.^(0.14)-(1-freqvector).^(0.09))/0.1596)
        
else
    phiminus1 = (freqvector.^(0.135) - (1-freqvector).^0.135)/0.1975
        
end;

frequencyvector is the array i have defined with my F calculations and phiminus1 will be my resulting array to be calculated.

The maths does not agree with my spreadsheet/hand calcs however and will not give the correct values, is someone able to figure out why?

Thank you for taking the time to read!

Answer 1

[Jan]

You did not show with which input you call your function. You mention, that there is a difference, but we cannot guess, which values you get, which you expect and which of them is wrong.

Therefore I can only guess boldly: Then term "freqvector" implies, that the value is a vector. Remember, that the condition of an if command must be a scalar. Therefore Matlab inserts this automatically:

if all(freqvector < 0.0028) & ~isempty(freqvector)
   ...
end

Is this wanted? I assume, you want to apply this function to the elements of the vector. Then you need either a loop:

phiminus = zeros(size(freqvector));   % Pre-allocation
for k = 1:numel(freqvector)
  if freqvector(k) < 0.0028
    phiminus1(k) = -(0.2 + ((1-freqvector(k)).^0.14 - (1 - (1-freqvector(k))).^0.09) / 0.1596);
  elseif freqvector(k) > 0.9972
    phiminus1(k) = 0.2 + (freqvector(k).^0.14 - (1 - freqvector(k)).^0.09) / 0.1596;
  else
    phiminus1(k) = (freqvector(k).^0.135 - (1-freqvector(k)).^0.09) / 0.1975;
  end
end

or you need a logical indexing:

phiminus     = zeros(size(freqvector));   % Pre-allocation
m1           = freqvector < 0.0028;
phiminus(m1) = -(0.2 + ((1-freqvector(m1)).^0.14 - (1 - (1 - freqvector(m1))).^0.09) / 0.1596);
m2           = freqvector > 0.9972;
phiminus(m2) = 0.2 + (freqvector(m2).^0.14 - (1 - freqvector(m2)).^0.09) / 0.1596;;
m3           = ~(m1 | m2);
phiminus(m3) = (freqvector(m3).^0.135 - (1 - freqvector(m3)).^0.09) / 0.1975;

Note: Spaces around operators improve the readability. Use parentheses if they are required or improve the clarity of the code. Too many parentheses look confusing.

[Drew Mullaney]

Jan youre a living legend.

Thank you so much. Very informative, easy to read and works perfectly.

[Jan]

Thanks, Drew. I'm happy, if I can help.