Feature-Weighting-for-DMI/DetectionKNNConfMat.m at digital-modulation-identification · sofiane-kharbech/Feature-Weighting-for-DMI · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78

%%%%% last update 26/05/2019

% vérifier les résultats avec normalisation

HOSth={

'   '  'B-PSK' 'Q-PSK' '8-PSK' '4-ASK' '8-ASK' '16-QAM'    ;
'M40'    1      -1       0       1.64    1.77    -0.67     ;
'M41'    1       0       0       1.64    1.77     0        ;
'M42'    1       1       1       1.64    1.77     1.32     ;
'M60'    1       0       0       2.92    3.62     0        ;
'M61'    1      -1       0       2.92    3.62    -1.32     ;
'M62'    1       0       0       2.92    3.62     0        ;
'M63'    1       1       1       2.92    3.62     1.96     ;
'M84'    1       1       1       5.25    7.92     3.12     ;
'C40'   -2      -1       0      -1.36   -1.42    -0.68     ;
'C41'   -2       0       0      -1.36   -1.42     0        ;
'C42'   -2      -1      -1      -1.36   -1.42    -0.68     ;
'C60'    16      0       0       8.32    7.19     0        ;
'C61'    16      4       0       8.32    7.19     2.08     ;
'C62'    16      0       0       8.32    7.19     0        ;
'C63'    16      4       4       8.32    7.19     2.08     ;

};

HOSth=cell2mat(HOSth(2:end,2:end));

lSNR=length(SNR);
lM=length(M);
lMMC=lM*lMC;

stat=zeros(4+Nt,lSNR*lMMC);

count=1;
for cptSNR=1:lSNR
    for cptM=1:lM
        for cptMC=1:lMC
            stat(1,count)=SNR(cptSNR);
            stat(2,count)=M(cptM);
            count=count+1;
        end
    end
end

% mat_HOS=mat_HOS(15:end,:,:);  %----> HOC
% HOSth=HOSth(9:end,:);         %----> HOC

% mat_HOS=mat_HOS([1 3 5 6 8 10 12 14:21],:,:);  %----> HOC and non-denoised HOM

mat_HOS=mat_HOS([1 2 4 6 7 9 11 13 15:21],:,:);  %----> HOC and denoised HOM


for cptNt=1:Nt

mat_HOS_cptNt=mat_HOS(:,:,cptNt);

[~,resUz]=min(pdist2(HOSth',mat_HOS_cptNt','euclidean'));
stat(2+cptNt,:)=resUz;

end

if Nt==1
stat(4,:)=stat(3,:);
else
stat(3+Nt,:)=mode(stat(3:2+Nt,:)); %nbr d'occurences
end
stat(end,:)=double(stat(2,:)==stat(end-1,:));

confMat = confusionmat(stat(2,:),stat(end-1,:));

save confMat

% plotConfMat(confMat, {'B-PSK' 'Q-PSK' '8-PSK' '4-ASK' '8-ASK' '16-QAM'})

% mean(stat(2,:)==stat(3,:))

! python3 confMat.py