Plot Data to find the decision boundary for logisitc regression.

Figure 1

I am using the Wisconsin Breast Cancer Diagnostic Data Set found on the UCI Machine Learning website. I am trying to implement logistic regression to classify the tumour as either malignant or benign, I changed the class to 0 for benign and 1 for malignant, for two features only, Marginal Adhesion and Clump Thickness. This is for an assignment at uni and we did not study much beyond logistic regression. I am aware there are other classification techniques out there but that is beyond the scope of my skills. I am trying to plot Maginal Adhesion against Clump Thickness however my graph ends up looking like the one above. I have used the following code that I found online.

This is the code to plot the data in a graph.

function plotData(X,y)
    figure; hold on
    pos = find(y==1); %Malignant
    neg = find(y==0); %Benign
    plot(X(pos,1), X(pos, 2), 'k+','LineWidth',5,'MarkerSize',5);
    plot(X(neg,1), X(neg,2), 'ko','MarkerFaceColor','b','MarkerSize',5);
    hold off;
end

Edit: I have included the code I used to plot the Decision Boundary, sorry I should have included that before too.

function plotDecisionBoundary(theta, X, y)

%Plot Data
plotData(X(:,2:3), y);
hold on

if size(X, 2) <= 3
    %Only need 2 points to define a line, so choose two endpoints
    plot_x = [min(X(:,2))-4, max(X(:,2))+4];
    
    %Calculate the decision boundary line
    plot_y = (-1./theta(3)).*(theta(2).*plot_x + theta(1));
    
    %Plot and adjust axes for better viewing
    plot(plot_x, plot_y)
    
    %Legend 
    legend('Benign','Malignant', 'Decision Boundary')
    axis([1,10, 1, 10])
else 
    %Here is the grid range
    u = linspace(-1, 1.5, 50);
    v = linspace(-1, 1.5, 50);
    
    z = zeros(length(u), length(v));
    %Evaluate z = theta*x over the grid
    for i = 1:length(u)
        for j = 1:length(v)
            z(i,j) = mapfeature(u(i), v(j))*theta;
        end 
    end 
    z = z'; %important to transpose z before contour
    
    %Plot z = 0
    %Notice you need to specify the range [0,0]
    countour(u, v, z, [0, 0], 'LineWidth', 2)
end 
hold off
end

To be honest, I am not too sure whether the code is right or not as I am still quite new to Matlab and am trying to get it working from watching videos online.

I want my graph to look like the following below; can anyone let me know what I am doing wrong?

Any help is appreciated. Thank you in advance.

Figure 2

[the cyclist]

Are you able to upload the X and y variables in a MAT file?

The figure you posted doesn't seem like it would come from the code you posted, because the code would output plus signs and circles as the plot markers (as you want), but your figure does not.

So ... I'm confused.

[Vivian Chow]

Do you mean I should upload the data set I used for the figure?

[the cyclist]

Yes, the ideal would be to upload everything needed for us to reproduce your plot, and see why you did not get what you expected.

[Vivian Chow]

Okay, I have attached the text file that I used, the first column is Marginal Adhesion, the second column is Clump Thickness and the third column is Class. I have not used the other columns.

[the cyclist]

I downloaded the data file, and it is immediately obvious that the data you posted are not the same as the data in the figure you are trying to reproduce. Specifically, the data you uploaded are integers, and the figure has values that are clearly non-integer.

So, it seems there is a problem with the data themselves.

[Vivian Chow]

Thank you so much for your help! This defintely cleared up my confusion as to why my graph always looked different.

Answer 1

[Gaurav Garg]

Hi Vivian,

You can look at the following link, where you can get more information on plotting model.

[Vivian Chow]

Thank you! I will defintely check it out