How do you change the color on a class in a legend?

I have only included the code that does the plotting:

figure; hold on

plot(Xred(1,:),Xred(2,:),'r.');

plot(Xblue(1,:),Xblue(2,:),'b.');

title('perceptron learning');

xlabel('x1');

ylabel('x2');

for z = 1:length(t_predict)

if t_predict(z)==1

plot(X(z,1),X(z,2),'ro')

end

if t_predict(z)==-1

plot(X(z,1),X(z,2),'bo')

end

end

legend('class +1','class -1','pred +1','pred -1');

I am plotting data points with a given label and then using sequential perceptron learning algorithms for a random set of weights to predict and plot those labels over the original data points. On the legend, pred -1 should be blue but it is showing as red. Is there a way to change it blue?

Answer 1

[Star Strider]

Try something like this —

figure; hold on

hp{1} = plot(Xred(1,:),Xred(2,:),'r.');

hp{2} = plot(Xblue(1,:),Xblue(2,:),'b.');

title('perceptron learning');

xlabel('x1');

ylabel('x2');

for z = 1:length(t_predict)

if t_predict(z)==1

hp{3} = plot(X(z,1),X(z,2),'ro');

end

if t_predict(z)==-1

hp{4} = plot(X(z,1),X(z,2),'bo');

end

end

legend([hp{1},hp{2},hp{3}(1),hp{4}(1)],'class +1','class -1','pred +1','pred -1');

The first two appear to be plotting correctly, however the last two have problems likely related to the how they are plotted. Selecting only the first element in each of the last two should work in situations such as these, since the legend call refers to them in the order they were plotted.

If this does not work, it will be necessary to see more of your code.

.

[Nicholas]

This worked out, thank you!

[Star Strider]

As always, my pleasure!