How do I exclude data coordinates from a preexisting array?

Hello!

I was wondering if anyone could help me, I am working on a project where I have a set of data inside of a matrix and I want to create a new matrix with that data using only points outside of an area defined by two functions.

For example, If I use data=rand(100,2) to create a 100X2 matrix and plot it, there is a seemingly random distribution of points. Let's say I plot y1= .5x+.45 and y2=.5x+.55. There will be some many of points inside of these two functions that I DON'T want to include in my data. How would I go about taking these out of the matrix and how would I go about plotting everything but those points?

I tried doing something like this:

>clear all
>close all

>x=linspace(-1,1,100);
>y=linspace(-1,1,100);

>limit1=.5*x+.45;
>limit2=.5*x+.55;

>data=rand(100,2);
>hold on
>plot(x,limit1);
>plot(x,limit2);
>xlim([0,1]);
ylim([0,1]);
>
>if data(:,2) < transpose(limit1)
    >scatter(data(:,1),data(:,2),'r','o');
    >if data(:,2) > transpose(limit2)
        >scatter(data(:,1),data(:,2),'r','o');
    >end
>else scatter(data(:,1),data(:,2),'*');
>end

Unfortunately it doesn't seem like it wants to plot the scatter plots correctly and I have no idea how to even start to remove the elements that don't satisfy the conditions from the array, any help would be greatly appreciated!

Answer 1

[James Tursa]

E.g., use logical indexing to isolate the points you want to plot.

g1 = data(:,2) < .5 * data(:,1) + .45; % Points less that lower limit
g2 = data(:,2) > .5 * data(:,1) + .55; % Points greater than upper limit
g = g1 | g2; % Union of the two point sets above
plot(data(g,1),data(g,2),'*'); % Plot only those points using g as 1st index

If you wanted to physically remove the unwanted points from your data set, then you could do this:

data = data(g,:); % Keep only the wanted points

[Cody Stein]

Oh my goodness, thank you so much! so if I unsuppress g, will it give me the matrix that has only the points that satisfy those conditions or is it going to give it to me with 1's and 0 stating true or false for each of those 100 rows?

[James Tursa]

g is a logical vector of 1's and 0's matching the conditions you set forth. ~g is a logical vector of the remaining points, which can also be used for indexing. E.g.,

plot(data(g,1),data(g,2),'*'); % Plot only those points meeting criteria
plot(data(~g,1),data(~g,2),'r*'); % Plot the other points in red

Your data is still your data ... it is not changed to 1's and 0's if that is what you are asking. The logical vector g is only used to select the data ... it doesn't change the values of the data.

[Cody Stein]

So if I wanted the new matrix that included only the data points that satisfied the conditions how would I go about getting that value? Would I multiply each column of data by the logical matrix 'g'? I might be doing that incorrectly because when I do that I get what appears to be a single answer.

[James Tursa]

Already answered above. The following line creates a new matrix consisting only of the data points that match your conditions (i.e., keeps all x-y pairs that match the conditions):

data_new = data(g,:); % Keep only the wanted points