/
maxRelMinRed.m
144 lines (131 loc) · 3.63 KB
/
maxRelMinRed.m
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
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
function [] = maxRelMinRed(data,class,str)
% Function to generate mRMR scores for a dataset
temp = zeros(size(class,1),1);
for i=1:size(data,1)
temp(i,1)=find(class(i,:),1);
end
class = temp;
clear temp;
bin = 10;
num = size(data,2);
redundancy = zeros(num);
relevance = zeros(1,num);
% data normalised
mean_feature = repmat(mean(data),size(data,1),1);
min_feature = repmat(min(data),size(data,1),1);
max_feature = repmat(max(data),size(data,1),1);
data = (data-mean_feature)./(max_feature-min_feature);
data = (data+1)/2; % from -1 to 1 => 0 to 1
clear max_feature min_feature mean_feature;
for i=1:num
for j=1:i-1
redundancy(i,j) = mutualInformationf(data(:,i),data(:,j),bin);
redundancy(j,i) = redundancy(i,j);
end
end
for i=1:num
relevance(1,i) = mutualInformationfC(data(:,i),class,bin);
end
save(strcat('Data/',str,'/redundancy.mat'),'redundancy');
save(strcat('Data/',str,'/relevancy.mat'),'relevance');
end
function [val] = mutualInformationf(xi,xj,bin)
val = entropy(xi,bin) + entropy(xj,bin) - jointEntropy(xi,xj,bin);
end
function [val] = mutualInformationfC(xi,class,bin)
val = entropy(xi,bin) + entropyClass(class) - jointEntropyClass(xi,class,bin);
end
function [val] = jointEntropy(xi,xj,bin)
if(isnan(xi))
val=0;
return;
end
if(isnan(xj))
val=0;
return;
end
edges = 0:(1.0/bin):1;
count = zeros(bin);
listi = discretize(xi',edges);
listj = discretize(xj',edges);
if (sum(isnan(listi)) || sum(isnan(listj)))
disp('error');
end
for i=1:size(xi,1)
count(listi(i),listj(i)) = count(listi(i),listj(i))+1;
end
count=count/(size(xi,1));
count=reshape(count,1,(bin^2));
val=0;
for i=1:size(count,2)
if count(1,i)~=0
val = val + (count(1,i)*log(count(1,i)));
end
end
val=-val;
end
function [val] = jointEntropyClass(xi,class,bin)
if(isnan(xi))
val=0;
return;
end
edges = 0:(1.0/bin):1;
listi = discretize(xi',edges);
if min(class)==0
class=class+1;
end
num = max(class);
count = zeros(bin,num);
if (sum(isnan(listi)) || sum(isnan(class)))
disp('error');
end
for i=1:size(xi,1)
count(listi(i),class(i)) = count(listi(i),class(i))+1;
end
count=count/(size(xi,1));
count=reshape(count,1,(bin*num));
val=0;
for i=1:size(count,2)
if count(1,i)~=0
val = val + (count(1,i)*log(count(1,i)));
end
end
val=-val;
end
function [val] = entropy(xi,bin)
edges = 0:(1.0/bin):1;
list = discretize(xi',edges);
if sum(isnan(list))
disp('error');
end
count = zeros(1,bin);
for i=1:bin
count(1,i) = sum(list==i);
end
count=count/size(xi,1);
val=0;
for i=1:size(count,2)
if count(1,i)~=0
val = val + (count(1,i)*log(count(1,i)));
end
end
val=-val;
end
function [val] = entropyClass(class)
if min(class) == 0
class=class+1;
end
bin=max(class);
count = zeros(1,bin);
for i=1:bin
count(1,i) = sum(class==i);
end
count=count/size(class,1);
val=0;
for i=1:size(count,2)
if count(1,i)~=0
val = val + (count(1,i)*log(count(1,i)));
end
end
val=-val;
end