how to make the regression line of lsline function colored
    18 visualizzazioni (ultimi 30 giorni)
  
       Mostra commenti meno recenti
    
    Jonas
 il 28 Feb 2011
  
    
    
    
    
    Commentato: Joris Bockhofer
 il 6 Lug 2023
            I have 3 scatter plots that are diplayed in the same graph in three different colors. I would like each of them to have a regression line of the same color. I can only get three grey regression line with the lsline command.
1 Commento
  Joris Bockhofer
 il 6 Lug 2023
				My recommendation would be to forget about lsline 
and simply calculate the linear regression yourself and plot the result separately with the color of the original plot. 
lsline has taken way more time out of my life than it should and i hope this finds anyone having trouble with coloring it.
My approach looks like this:
rng default  % For reproducibility
x = 1:10;
y1 = x + randn(1,10); 
y2 = 2*x + randn(1,10); % same size as x
figure
ax1 = gca;
hold(ax1, 'on');
xData = [x; x];
yData = [y1;y2];
legendStrArray = string(); % creat a string array to hold label information
for i = 1:numel(xData(:,1))
    dataX = [xData(i,:)];
    dataY = [yData(i,:)];
    pl1 = scatter(ax1, dataX, dataY); % plot original data    
    colorRGB = pl1.CData; % get color of plot
    % Fit a linear model
    model = fitlm(dataX, dataY);
    % Get the fitted line data
    xFit = linspace(min(dataX), max(dataX), 100);
    yFit = feval(model, xFit);
    h = plot(ax1, xFit, yFit, 'Color', colorRGB); % plot regression with color
    % add to legend with end+1 strategy for when you dont want to set up a dedicated counter
    legendStrArray(1,end+1) = "Datasheet: " + string(i); % legend for data
    legendStrArray(1,end+1) = "Datasheet: " + string(i) + ", Regression"; % legend for it's regression
      % otherwise it mess up the number of plots vs the number of legend entries
end
legendStrArray = legendStrArray(1,2:end);% take off the first empty entry from the legend array bc of end+1
legend(ax1,legendStrArray, 'Location', 'best'); % have a working legend
hold(ax1, 'off');
Why complicated when it could be simple (or was it the other way around?)  - Hope this saves you the headache!
Risposta accettata
  Oleg Komarov
      
      
 il 28 Feb 2011
        When you add the regression line call it like this:
h = lsline;
set(h(1),'color','r')
set(h(2),...
Oleg
3 Commenti
  Joris Bockhofer
 il 6 Lug 2023
				Honestly this should be made way more obvious, i spend way too long just trying to set color because the way in the documentation wasnt working so i had to do : [h1.Color] = deal(color); because it kept giving me assignment errors, and then i had to find out that lsline automatically plots a line for each held scatterplot, so setting the color from inside the same loop as you are plotting the scatterplot doesnt seem possible. you'd have to create a second loop for iterating over the handles.... this is just top notch
Più risposte (0)
Vedere anche
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!




