% *************************************************************************

% CS4019 - Computational Pattern Recognition
% Science Faculty
% University of Colombo

% Assignment I (Feature Extraction)

% Name: THARINDRA DHARMAJITH WIJAYATILLEKE
% Registration No: 2005/s/9771
% Index No: s 8323

% *************************************************************************


clear;
clc;
close all;

dataFile=load('2005s9771-data.txt'); 



mcg(:,1)=dataFile(:,2); %mcg: McGeoch's method for signal sequence recognition.                                               
gvh(:,1)=dataFile(:,3); %gvh: von Heijne's method for signal sequence recognition.                                    
alm(:,1)=dataFile(:,4); %alm: Score of the ALOM membrane spanning region prediction program.                               
mit(:,1)=dataFile(:,5); %mit: Score of discriminant analysis of the amino acid content of
	                         %the N-terminal region (20 residues long) of mitochondrial and 
                            % non-mitochondrial proteins.
      
erl(:,1)=dataFile(:,6); %erl: Presence of "HDEL" substring (thought to act as a signal for
	                         %retention in the endoplasmic reticulum lumen). Binary attribute.
           
pox(:,1)=dataFile(:,7); %pox: Peroxisomal targeting signal in the C-terminus.
            
vac(:,1)=dataFile(:,8); %vac: Score of discriminant analysis of the amino acid content of
                             %vacuolar and extracellular proteins.                                 
nuc(:,1)=dataFile(:,9); %nuc: Score of discriminant analysis of nuclear localization signals
	                         %of nuclear and non-nuclear proteins.



   
t=1:1136;

%Attributes
   
figure(1)
plot(t(1:463),mcg(1:463),'r*');
hold on;
plot(t(1:429),mcg(464:892),'b*');
hold on;
plot(t(1:244),mcg(893:1136),'g*');
hold on;
title('mcg');
hold off;


figure(2)
plot(t(1:463),gvh(1:463),'r*');
hold on;
plot(t(1:429),gvh(464:892),'b*');
hold on;
plot(t(1:244),gvh(893:1136),'g*');
hold on;
title('gvh');
hold off;


figure(3)
plot(t(1:463),alm(1:463),'r*');
hold on;
plot(t(1:429),alm(464:892),'b*');
hold on;
plot(t(1:244),alm(893:1136),'g*');
hold on;
title('alm');
hold off;


figure(4)
plot(t(1:463),mit(1:463),'r*');
hold on;
plot(t(1:429),mit(464:892),'b*');
hold on;
plot(t(1:244),mit(893:1136),'g*');
hold on;
title('mit)');
hold off;


figure(5)
plot(t(1:463),erl(1:463),'r*');
hold on;
plot(t(1:429),erl(464:892),'b*');
hold on;
plot(t(1:244),erl(893:1136),'g*');
hold on;
title('erl.');
hold off;


figure(6)
plot(t(1:463),pox(1:463),'r*');
hold on;
plot(t(1:429),pox(464:892),'b*');
hold on;
plot(t(1:244),pox(893:1136),'g*');
hold on;
title('pox.');
hold off;


figure(7)
plot(t(1:463),vac(1:463),'r*');
hold on;
plot(t(1:429),vac(464:892),'b*');
hold on;
plot(t(1:244),vac(893:1136),'g*');
hold on;
title('vac');
hold off;


figure(8)
plot(t(1:463),nuc(1:463),'r*');
hold on;
plot(t(1:429),nuc(464:892),'b*');
hold on;
plot(t(1:244),nuc(893:1136),'g*');
hold on;
title('nuc');
hold off;

%According to figure(5)and figure(6) attributes erl and pox are less
%important



%Features


figure(9)
plot(t(1:463),mcg(1:463).*nuc(1:463),'r*');
hold on;
plot(t(1:429),mcg(464:892).*nuc(464:892),'b*');
hold on;
plot(t(1:244),mcg(893:1136).*nuc(893:1136),'g*');
hold on;
title('mcg and nuc');
hold off;

figure(10)
plot(t(1:463),nuc(1:463)./mcg(1:463),'r*');
hold on;
plot(t(1:429),nuc(464:892)./mcg(464:892),'b*');
hold on;
plot(t(1:244),nuc(893:1136)./mcg(893:1136),'g*');
hold on;
title('mcg and nuc');
hold off;



figure(11)
plot(t(1:463),mcg(1:463).*gvh(1:463),'r*');
hold on;
plot(t(1:429),mcg(464:892).*gvh(464:892),'b*');
hold on;
plot(t(1:244),mcg(893:1136).*gvh(893:1136),'g*');
hold on;
title('mcg and gvh');
hold off;

figure(12)
plot(t(1:463),mcg(1:463)./gvh(1:463),'r*');
hold on;
plot(t(1:429),mcg(464:892)./gvh(464:892),'b*');
hold on;
plot(t(1:244),mcg(893:1136)./gvh(893:1136),'g*');
hold on;
title('mcg and gvh');
hold off;



figure(13)
plot(t(1:463),mcg(1:463).*alm(1:463),'r*');
hold on;
plot(t(1:429),mcg(464:892).*alm(464:892),'b*');
hold on;
plot(t(1:244),mcg(893:1136).*alm(893:1136),'g*');
hold on;
title('mcg and alm');
hold off;

figure(14)
plot(t(1:463),alm(1:463)./mcg(1:463),'r*');
hold on;
plot(t(1:429),alm(464:892)./mcg(464:892),'b*');
hold on;
plot(t(1:244),alm(893:1136)./mcg(893:1136),'g*');
hold on;
title('mcg and alm');
hold off;



figure(15)
plot(t(1:463),mcg(1:463).*mit(1:463),'r*');
hold on;
plot(t(1:429),mcg(464:892).*mit(464:892),'b*');
hold on;
plot(t(1:244),mcg(893:1136).*mit(893:1136),'g*');
hold on;
title('mcg and mit');
hold off;

figure(16)
plot(t(1:463),mit(1:463)./mcg(1:463),'r*');
hold on;
plot(t(1:429),mit(464:892)./mcg(464:892),'b*');
hold on;
plot(t(1:244),mit(893:1136)./mcg(893:1136),'g*');
hold on;
title('mcg and mit');
hold off;




figure(17)
plot(t(1:463),mcg(1:463).*vac(1:463),'r*');
hold on;
plot(t(1:429),mcg(464:892).*vac(464:892),'b*');
hold on;
plot(t(1:244),mcg(893:1136).*vac(893:1136),'g*');
hold on;
title('mcg and vac');
hold off;

figure(18)
plot(t(1:463),vac(1:463)./mcg(1:463),'r*');
hold on;
plot(t(1:429),vac(464:892)./mcg(464:892),'b*');
hold on;
plot(t(1:244),vac(893:1136)./mcg(893:1136),'g*');
hold on;
title('mcg and vac');
hold off;



figure(19)
plot(t(1:463),gvh(1:463).*alm(1:463),'r*');
hold on;
plot(t(1:429),gvh(464:892).*alm(464:892),'b*');
hold on;
plot(t(1:244),gvh(893:1136).*alm(893:1136),'g*');
hold on;
title('gvh and alm');
hold off;

figure(20)
plot(t(1:463),alm(1:463)./gvh(1:463),'r*');
hold on;
plot(t(1:429),alm(464:892)./gvh(464:892),'b*');
hold on;
plot(t(1:244),alm(893:1136)./gvh(893:1136),'g*');
hold on;
title('gvh and alm');
hold off;



figure(21)
plot(t(1:463),gvh(1:463).*mit(1:463),'r*');
hold on;
plot(t(1:429),gvh(464:892).*mit(464:892),'b*');
hold on;
plot(t(1:244),gvh(893:1136).*mit(893:1136),'g*');
hold on;
title('gvh and mit');
hold off;

figure(22)
plot(t(1:463),mit(1:463)./gvh(1:463),'r*');
hold on;
plot(t(1:429),mit(464:892)./gvh(464:892),'b*');
hold on;
plot(t(1:244),mit(893:1136)./gvh(893:1136),'g*');
hold on;
title('gvh and mit');
hold off;



figure(23)
plot(t(1:463),gvh(1:463).*vac(1:463),'r*');
hold on;
plot(t(1:429),gvh(464:892).*vac(464:892),'b*');
hold on;
plot(t(1:244),gvh(893:1136).*vac(893:1136),'g*');
hold on;
title('gvh and vac');
hold off;

figure(24)
plot(t(1:463),vac(1:463)./gvh(1:463),'r*');
hold on;
plot(t(1:429),vac(464:892)./gvh(464:892),'b*');
hold on;
plot(t(1:244),vac(893:1136)./gvh(893:1136),'g*');
hold on;
title('gvh and vac');
hold off;



figure(25)
plot(t(1:463),gvh(1:463).*nuc(1:463),'r*');
hold on;
plot(t(1:429),gvh(464:892).*nuc(464:892),'b*');
hold on;
plot(t(1:244),gvh(893:1136).*nuc(893:1136),'g*');
hold on;
title('gvh and nuc');
hold off;

figure(26)
plot(t(1:463),nuc(1:463)./gvh(1:463),'r*');
hold on;
plot(t(1:429),nuc(464:892)./gvh(464:892),'b*');
hold on;
plot(t(1:244),nuc(893:1136)./gvh(893:1136),'g*');
hold on;
title('gvh and nuc');
hold off;



figure(27)
plot(t(1:463),alm(1:463).*vac(1:463),'r*');
hold on;
plot(t(1:429),alm(464:892).*vac(464:892),'b*');
hold on;
plot(t(1:244),alm(893:1136).*vac(893:1136),'g*');
hold on;
title('alm and vac');
hold off;

figure(28)
plot(t(1:463),vac(1:463)./alm(1:463),'r*');
hold on;
plot(t(1:429),vac(464:892)./alm(464:892),'b*');
hold on;
plot(t(1:244),vac(893:1136)./alm(893:1136),'g*');
hold on;
title('alm and vac');
hold off;



figure(29)
plot(t(1:463),alm(1:463).*nuc(1:463),'r*');
hold on;
plot(t(1:429),alm(464:892).*nuc(464:892),'b*');
hold on;
plot(t(1:244),alm(893:1136).*nuc(893:1136),'g*');
hold on;
title('alm and nuc');
hold off;

figure(30)
plot(t(1:463),nuc(1:463)./alm(1:463),'r*');
hold on;
plot(t(1:429),nuc(464:892)./alm(464:892),'b*');
hold on;
plot(t(1:244),nuc(893:1136)./alm(893:1136),'g*');
hold on;
title('alm and nuc');
hold off;

%features are investigated by considering these graphs