Scribd
Upload
93 views3 pages

Pattern Recognition Home Assignment 2

This document presents a MATLAB code that aims to classify the Fisher Iris data set into three classes (setosa, versicolor, virginica) using decision boundaries based on petal length and petal width. The code loads the Fisher Iris data, calculates the mean, standard deviation and covariance for each class, defines decision boundaries using Gaussian distributions, and plots the results showing clear separation of the three classes.

Uploaded by

Sagar Gangadhar
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
93 views3 pages

Pattern Recognition Home Assignment 2

This document presents a MATLAB code that aims to classify the Fisher Iris data set into three classes (setosa, versicolor, virginica) using decision boundaries based on petal length and petal width. The code loads the Fisher Iris data, calculates the mean, standard deviation and covariance for each class, defines decision boundaries using Gaussian distributions, and plots the results showing clear separation of the three classes.

Uploaded by

Sagar Gangadhar
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

PATTERN RECOGNITION HOME ASSIGNMENT 2

Submitted by: Sagar G Yadav [CB107EC516], Anand Chandran [CB107EC006],


Arun Kumar [CB107EC017]

AIM:
 To classify the Fisher Iris Standard Data Set into three classes namely “setosa”,
“versicolor” and “virginica” using decision boundaries.
 The parameters: Petal Length and Petal Width is used to classify the data set.

MATLAB CODE:

clc;
clear all;
close all;

load fisheriris; % Load Fisher Iris Data %


%implicit columns: Sepal Length, Sepal Width, Petal Length, Petal Width%

setosa_indices = strcmp('setosa',species);
setosa = meas(setosa_indices,:);

versicolor_indices = strcmp('versicolor',species);
versicolor = meas(versicolor_indices,:);

virginica_indices = strcmp('virginica',species);
virginica = meas(virginica_indices,:);

setosaOBS = setosa(:,3:4);
versicolorOBS = versicolor(:,3:4);
virginicaOBS = virginica(:,3:4);

r_setosa = corrcoef(setosaOBS);
r_versicolor = corrcoef(versicolorOBS);
r_virginica = corrcoef(virginicaOBS);

sig_setosa = [std2(setosaOBS(:,1)), std2(setosaOBS(:,2))];


sig_versicolor = [std2(versicolorOBS(:,1)), std2(versicolorOBS(:,2))];
sig_virginica = [std2(virginicaOBS(:,1)), std2(virginicaOBS(:,2))];

cov_setosa = corr2cov(sig_setosa, r_setosa);


cov_versicolor = corr2cov(sig_versicolor, r_versicolor);
cov_virginica = corr2cov(sig_virginica, r_virginica);

cov = cat(3,cov_setosa, cov_versicolor, cov_virginica);

u_setosa = [mean(setosaOBS(:,1)), mean(setosaOBS(:,2))];


u_versicolor = [mean(versicolorOBS(:,1)), mean(versicolorOBS(:,2))];
u_virginica = [mean(virginicaOBS(:,1)), mean(virginicaOBS(:,2))];
centres = [u_setosa; u_versicolor; u_virginica];
priors = [(1/3), (1/3), (1/3)];
nsamples = 50;

gridres = 100;
x_min = 0;
x_max = 7;
y_min = 0;
y_max = 3;
xspace = linspace(x_min, x_max, gridres);
yspace = linspace(y_min, y_max, gridres);

[X Y] = meshgrid(xspace, yspace);
gridpoints = [X(:) Y(:)];
npoints = size(gridpoints, 1);

px_C = zeros(npoints, 2);


normalfact = (2*pi);

for i = 1:3
dist = gridpoints - (ones(npoints,1)*centres(i,:));
Q = chol(cov(:,:,i));
term = dist*inv(Q);
px_C(:,i) = exp(-0.5*sum(term .* term, 2))./(normalfact*prod(diag(Q)));
end

postnum = (ones(npoints, 1)*priors).*px_C;


px = sum(postnum, 2);
posteriors = postnum ./ (px*ones(1,3));

p1_x = reshape(posteriors(:,1), size(X));


p2_x = reshape(posteriors(:,2), size(X));
p3_x = reshape(posteriors(:,3), size(X));
px_1 = reshape(px_C(:,1), size(X));
px_2 = reshape(px_C(:,2), size(X));
px_3 = reshape(px_C(:,3), size(X));

contour(xspace, yspace, px_1);


hold on;
contour(xspace, yspace, px_2);
hold on;
contour(xspace, yspace, px_3);
hold on;
contour(xspace, yspace, p2_x, [(1/3) (1/3) (1/3)], '-.r*');
hold on;
contour(xspace, yspace, p3_x, [(1/3) (1/3) (1/3)], '--go');
hold on;
contour(xspace, yspace, p1_x, [(1/3) (1/3) (1/3)], ':bs');
hold on;
plot(setosaOBS(:,1), setosaOBS(:,2),'k.');
hold on;
plot(versicolorOBS(:,1), versicolorOBS(:,2),'k*');
hold on;
plot(virginicaOBS(:,1), virginicaOBS(:,2),'ko');

hold on;
annotation('textbox',...
[0.167071428571429 0.20952380952381 0.133928571428571
0.0642857142857143],...
'String',{'SETOSA'},...
'FontWeight','bold',...
'FontSize',9,...
'FontName','Arial',...
'FontAngle','italic',...
'LineStyle','none');

annotation('textbox',...
[0.4135 0.266666666666668 0.189285714285714 0.0642857142857143],...
'String',{'VERSICOLOR'},...
'FontWeight','bold',...
'FontSize',9,...
'FontName','Arial',...
'FontAngle','italic',...
'LineStyle','none');

annotation('textbox',...
[0.595642857142856 0.785714285714289 0.158928571428571
0.0642857142857143],...
'String',{'VIRGINICA'},...
'FontWeight','bold',...
'FontSize',9,...
'FontName','Arial',...
'FontAngle','italic',...
'LineStyle','none');

xlabel('PETAL LENGTH -->','fontsize', 8);


ylabel('PETAL WIDTH -->','fontsize', 8);
title('FISHER IRIS DATA CLASSIFICATION USING DECISION
BOUNDARIES','fontsize', 9,'fontweight', 'b');

RESULT:

FISHER IRIS DATA CLASSIFICATION USING DECISION BOUNDARIES


3

VIRGINICA
2.5

2
PETAL WIDTH -->

1.5

VERSICOLOR

0.5 SETOSA

0
0 1 2 3 4 5 6 7
PETAL LENGTH -->

You might also like