How can I change input of function from file to cell array?

Hi,

I have this function dfaedit_2('filename.csv',0,0,0) and it only takes files as input. I have a cell array with a size of 19x21.

I would like to change the input into a cell array.

When I run it however I get an error:

Error using cumsum

Invalid data type. First input argument must be numeric or logical.

Error in dfaedit (line 44)

Sum = cumsum(x);

Help is much appreciated!

Answer 1

[David Hill]

Either delete or comment out lines below and change input to data. Verify cell2mat(of your cell array) gives what is expected. Or provide a sample of your input cell array and previous data file.

data=yourCellarray;
H=dfaedit(data,1,1,1);%call function

function  [H]=dfaedit(data,plot_flag, outfile_flag, out_command_flag)
format short g
%x = file_name;
%fid = fopen(file_name);
%data = textscan(fid,'%f');
%fid = fclose(fid);
x = cell2mat(data);

[lil brain]

Hi @David Hill,

thanks for the quick response! When I call the function

H=dfaedit_2(p_windows,1,1,1)

I get the error:

Error using cumsum

Invalid data type. First input argument must be numeric or logical.

Error in dfaedit (line 44)

Sum = cumsum(x);

When I run

x = cell2mat(data)

I get the error:

Error using cat

Dimensions of arrays being concatenated are not consistent.

Error in cell2mat (line 78)

m = cat(1,m{:});

So does that mean I cant use cell2mat?

Thanks for the help!

[David Hill]

function  [H]=dfaedit(data,plot_flag, out_command_flag)

format short g

x = data{1};

numberpoints = length(x);

MinBox = 4;                 % minbox #of points in a box
MaxBox  = .25*numberpoints; % maxbox set 1/4 the data length
BoxSizeDensity = 4;
LogScaleFactor = power(2.0, 1.0/BoxSizeDensity);
index = 1:numberpoints; 
index = reshape(index, length(index), 1);  %change row to column vector

% Preliminary calculations
Sum = cumsum(x);
SumOfSumsSquared = cumsum(Sum.*Sum);
SumOfSums = cumsum(Sum);
SSc = cumsum(index.*Sum);

% now find best fit lines and find fluctuation about the line 
% loop for each box size, from MinBox to MaxBox
iteration = 1;
BoxSize = MinBox;
TempBoxSize = MinBox;
while BoxSize <= MaxBox
    s = 0.0;
    r1 = 1./(BoxSize + 1.0);
    Det = 12.0*r1*r1*r1/(1.0 - r1*r1);
    vv = BoxSize*(2.0*BoxSize + 1.0)/6.0;
    
    for j = 2:numberpoints - BoxSize
        dhh = SumOfSumsSquared(j + BoxSize) - SumOfSumsSquared(j-1);
        dhu = SumOfSums(j + BoxSize) - SumOfSums(j-1);
        dhv = SSc(j + BoxSize) - SSc(j-1) - dhu*j;
        
        s = s + dhh - (dhv*dhv +dhu*dhu*vv - dhv*dhu*BoxSize)*Det;
        j = j + 1;
    end
    
    den = (numberpoints - BoxSize)*(BoxSize + 1.0);
    Fluctuation = sqrt(s/den);
    
    log_points_in_box(iteration,1) = log10(BoxSize);  
    log_Q(iteration,1) = log10(Fluctuation);
    show_results(iteration,:) = [ iteration  BoxSize  Fluctuation  log_points_in_box(iteration,1)  log_Q(iteration,1) ];
    iteration = iteration + 1;
    
    % update the box size 
    TempBoxSize = TempBoxSize*LogScaleFactor;
    while round(TempBoxSize) <  BoxSize + 1.0
        TempBoxSize = TempBoxSize*LogScaleFactor;
    end
    BoxSize = round(TempBoxSize);
end

% got all boxes; now calculate H via trendline
r_trend = corrcoef(log_points_in_box, log_Q);
coefs = polyfit(log_points_in_box, log_Q,1);
r_line = polyval(coefs,log_points_in_box);

% calculate dimension and Hurst
H = coefs(1);
D = 2-H;

%display results in command window
if out_command_flag == 1
    disp('         iteration    points       Q       Log10(pnts)   Log10(Q)')
    disp(show_results)
end

% display results in a figure
if plot_flag == 1
    figure
    h = axes('Position', [0 0 1 1], 'Visible', 'off');
    axes('Position',[.1 .2 .75 .75])
    plot(log_points_in_box, log_Q,'b+')
    hold on
    plot(log_points_in_box,r_line, 'k')
    ylabel('log10(Q)');
    xlabel('log10(Points in Subset)')
    set(gcf, 'CurrentAxes', h);
    str(1) = {[sprintf('%2.5f',coefs(1)),'x + ', sprintf('%2.5f',coefs(2)), ' = y,   r^2 = ', sprintf('%2.2f',r_trend(1,2)^2), ',  n = ', sprintf('%d',length(log_Q)) ]};
    text(.1, .02, str, 'FontSize', 10, 'Color', 'r');
    str(1) = {['H = ', sprintf('%2.3f',H) '  D = ', sprintf('%2.3f',D) ]};
    text(.5, .02, str, 'FontSize', 10, 'Color', 'r');
end

end

Call function

h=dfaedit(p_windows11,1,1);

[lil brain]

Hi @David Hill,

thanks this worked. Very grateful for your answer.

I was wondering however, how do I get an result (h) for every column in each of the cells. I was hoping to output a cell array with an h for each column. Apologies if I didnt mention that at the top when asking the question. Could you help me with this?

[lil brain]

I tried using cellfun by calling the function like this:

wrapper = @(x)dfaedit(p_windows11,0,0);
h= cellfun(wrapper, p_windows, 'UniformOutput',false);

But this only yields an 19x21 cell array where each cell is the same result (see attachment).

[David Hill]

Not sure if you saw this line:

x = data{1};

I ran out of time and quickly put out my response. In order to find the h values for all parts of the cell array, you will need to change that line to cycle through the whole cell array.

for k=1:numel(data)
   x=data{k};
   % put the rest of the code in here
end

Like:

function  [H]=dfaedit(data,plot_flag, out_command_flag)

format short g
for k=1:numel(data)
x = data{k};

numberpoints = length(x);

MinBox = 4;                 % minbox #of points in a box
MaxBox  = .25*numberpoints; % maxbox set 1/4 the data length
BoxSizeDensity = 4;
LogScaleFactor = power(2.0, 1.0/BoxSizeDensity);
index = 1:numberpoints; 
index = reshape(index, length(index), 1);  %change row to column vector

% Preliminary calculations
Sum = cumsum(x);
SumOfSumsSquared = cumsum(Sum.*Sum);
SumOfSums = cumsum(Sum);
SSc = cumsum(index.*Sum);

% now find best fit lines and find fluctuation about the line 
% loop for each box size, from MinBox to MaxBox
iteration = 1;
BoxSize = MinBox;
TempBoxSize = MinBox;
while BoxSize <= MaxBox
    s = 0.0;
    r1 = 1./(BoxSize + 1.0);
    Det = 12.0*r1*r1*r1/(1.0 - r1*r1);
    vv = BoxSize*(2.0*BoxSize + 1.0)/6.0;
    
    for j = 2:numberpoints - BoxSize
        dhh = SumOfSumsSquared(j + BoxSize) - SumOfSumsSquared(j-1);
        dhu = SumOfSums(j + BoxSize) - SumOfSums(j-1);
        dhv = SSc(j + BoxSize) - SSc(j-1) - dhu*j;
        
        s = s + dhh - (dhv*dhv +dhu*dhu*vv - dhv*dhu*BoxSize)*Det;
        j = j + 1;
    end
    
    den = (numberpoints - BoxSize)*(BoxSize + 1.0);
    Fluctuation = sqrt(s/den);
    
    log_points_in_box(iteration,1) = log10(BoxSize);  
    log_Q(iteration,1) = log10(Fluctuation);
    show_results(iteration,:) = [ iteration  BoxSize  Fluctuation  log_points_in_box(iteration,1)  log_Q(iteration,1) ];
    iteration = iteration + 1;
    
    % update the box size 
    TempBoxSize = TempBoxSize*LogScaleFactor;
    while round(TempBoxSize) <  BoxSize + 1.0
        TempBoxSize = TempBoxSize*LogScaleFactor;
    end
    BoxSize = round(TempBoxSize);
end

% got all boxes; now calculate H via trendline
r_trend = corrcoef(log_points_in_box, log_Q);
coefs = polyfit(log_points_in_box, log_Q,1);
r_line = polyval(coefs,log_points_in_box);

% calculate dimension and Hurst
H(k) = coefs(1);
D = 2-H(k);

%display results in command window
if out_command_flag == 1
    disp('         iteration    points       Q       Log10(pnts)   Log10(Q)')
    disp(show_results)
end

% display results in a figure
if plot_flag == 1
    figure
    h = axes('Position', [0 0 1 1], 'Visible', 'off');
    axes('Position',[.1 .2 .75 .75])
    plot(log_points_in_box, log_Q,'b+')
    hold on
    plot(log_points_in_box,r_line, 'k')
    ylabel('log10(Q)');
    xlabel('log10(Points in Subset)')
    set(gcf, 'CurrentAxes', h);
    str(1) = {[sprintf('%2.5f',coefs(1)),'x + ', sprintf('%2.5f',coefs(2)), ' = y,   r^2 = ', sprintf('%2.2f',r_trend(1,2)^2), ',  n = ', sprintf('%d',length(log_Q)) ]};
    text(.1, .02, str, 'FontSize', 10, 'Color', 'r');
    str(1) = {['H = ', sprintf('%2.3f',H(k)) '  D = ', sprintf('%2.3f',D) ]};
    text(.5, .02, str, 'FontSize', 10, 'Color', 'r');
end
end
end

[lil brain]

Hi @David Hill,

sorry again for not explaining this correctly. When I run the code I get a result for each cell. However what I was looking for is a result for each column of each matrix within each cell.

For example, the first cell in p_windows11 has a size of 512x47. Therefor, I would need an output where the first cell in the new cell array (h) contains a matrix with 47 columns (1x47) or 47 results respectively.

I hope this is more clear. Again, support is very much appreciated!

[David Hill]

function  [H]=dfaedit(data)
format short g
for k=1:numel(data)
 y = data{k};
 h=[];
 for kk=1:size(y,2)
    x=y(:,kk);
    numberpoints = length(x);
    MinBox = 4;                 % minbox #of points in a box
    MaxBox  = .25*numberpoints; % maxbox set 1/4 the data length
    BoxSizeDensity = 4;
    LogScaleFactor = power(2.0, 1.0/BoxSizeDensity);
    index = 1:numberpoints; 
    index = reshape(index, length(index), 1);  %change row to column vector
    % Preliminary calculations
    Sum = cumsum(x);
    SumOfSumsSquared = cumsum(Sum.*Sum);
    SumOfSums = cumsum(Sum);
    SSc = cumsum(index.*Sum);
    % now find best fit lines and find fluctuation about the line 
    % loop for each box size, from MinBox to MaxBox
    iteration = 1;
    BoxSize = MinBox;
    TempBoxSize = MinBox;
    while BoxSize <= MaxBox
        s = 0.0;
        r1 = 1./(BoxSize + 1.0);
        Det = 12.0*r1*r1*r1/(1.0 - r1*r1);
        vv = BoxSize*(2.0*BoxSize + 1.0)/6.0;
    
        for j = 2:numberpoints - BoxSize
            dhh = SumOfSumsSquared(j + BoxSize) - SumOfSumsSquared(j-1);
            dhu = SumOfSums(j + BoxSize) - SumOfSums(j-1);
            dhv = SSc(j + BoxSize) - SSc(j-1) - dhu*j;
            s = s + dhh - (dhv*dhv +dhu*dhu*vv - dhv*dhu*BoxSize)*Det;
            j = j + 1;
        end
    
        den = (numberpoints - BoxSize)*(BoxSize + 1.0);
        Fluctuation = sqrt(s/den);
    
        log_points_in_box(iteration,1) = log10(BoxSize);  
        log_Q(iteration,1) = log10(Fluctuation);
        show_results(iteration,:) = [ iteration  BoxSize  Fluctuation  log_points_in_box(iteration,1)  log_Q(iteration,1) ];
        iteration = iteration + 1;
    
        % update the box size 
        TempBoxSize = TempBoxSize*LogScaleFactor;
        while round(TempBoxSize) <  BoxSize + 1.0
            TempBoxSize = TempBoxSize*LogScaleFactor;
        end
        BoxSize = round(TempBoxSize);
    end
    % got all boxes; now calculate H via trendline
    r_trend = corrcoef(log_points_in_box, log_Q);
    coefs = polyfit(log_points_in_box, log_Q,1);
    r_line = polyval(coefs,log_points_in_box);
    % calculate dimension and Hurst
    h(kk) = coefs(1);
    D = 2-h(kk);
 end
 H{k}=h;
end
end

Answer 2

[lil brain]

Beautiful! Many thanks!

Sorry one last thing, I promise! When running this code the resulting cells are saved in a single row rather than in the same shape that the cell array p_windows11 was originally in.

I am looking to have the cells in H saved in the same sequence as in p_windows11. Meaning that all the results from the cells in the first column are saved column-wise first in H as well.

Example:

H

1x47 = H{1,1}

1x43 = H{2,1}

Sorry for my poor experience in matlab!

[David Hill]

function  [H]=dfaedit(data)
format short g
H=cell(size(data));%just add this line at the beginning

[lil brain]

Perfect! This is exctly what I needed! If my project gets published, you're getting special thanks!