-
Notifications
You must be signed in to change notification settings - Fork 0
/
components_multi.m
253 lines (207 loc) · 7.15 KB
/
components_multi.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
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
function [res] = components_multi(stack_all,tr_labels,nfolds,hsize,pname,stack_allcompl)
ncompoini=2;
Tama=size(stack_all);
if nargin>5, tama2=size(stack_allcompl); P=Tama(1)+tama2(1); else stack_allcompl=[]; P=Tama(1); end
[train,test] = crossvalind('HoldOut',P,hsize);
okopts={'svm','bayes'};
%pname = 'bayes';% 'svm';%
labels=(tr_labels>0);
res=struct('TestValues',cell(1,1),'TrainValues',cell(1,1),'map',cell(1,1));
labels_train=labels(train);
labels_test=labels(test);
%% Entrenamiento de las componentes
fprintf('Empezando con el grupo ')
%lista=listar(stack_all,1,2);
load('listado.mat');
if Tama(2)>100
lista=listado.tama121x145x121.aal;
else
lista=listado.tama69x95x79.aal;
end
numOfcomp=size(lista,2);
for i=ncompoini:numOfcomp
fprintf([ 'Calculando la precision de la componente ' num2str(i) '/' num2str(numOfcomp) ]);
try
if nargin>5
component_part1=stack_all(:,lista{i});
component_part2=stack_allcompl(:,lista{i});
component=[component_part1;component_part2];
stack_train=component(train,:);
else
stack_train=stack_all(train,lista{i});
end
train_values(i)=component_accuracy(stack_train,labels_train,nfolds,pname);
%train_values(i)=component_minSV(stack_train,labels_train,pname);
catch
train_values(i)=NaN;
end
fprintf('completado! \n')
end
%% Validación de las componentes
fprintf('Validando el mapa de precision..')
cl = find(strncmpi(pname, okopts,numel(pname)));
switch cl
case 1
for i=ncompoini:numOfcomp
if nargin>5
component_part1=stack_all(:,lista{i});
component_part2=stack_allcompl(:,lista{i});
component=[component_part1;component_part2];
test_data=component(test,:);
else
test_data=stack_all(test,lista{i});
end
classes = svmclassify(train_values(i).trad,test_data);
cp=classperf(labels_test, classes);
test_values(i).cp=cp;
end
case 2
for i=ncompoini:numOfcomp
if nargin>5
component_part1=stack_all(:,lista{i});
component_part2=stack_allcompl(:,lista{i});
component=[component_part1;component_part2];
test_data=component(test,:);
else
test_data=stack_all(test,lista{i});
end
classes = predict(train_values(i).trad,test_data);
cp=classperf(labels_test, double(classes));
test_values(i).cp=cp;
end
end
fprintf('Completado! \n')
%% Elaborar mapa de precision
fprintf('Elaborando mapa de precision..')
Z=Tama(2);Y=Tama(3);X=Tama(4);
pmap=zeros(Z,Y,X);
spmap=zeros(Z,Y,X);
semap=zeros(Z,Y,X);
counter=zeros(Z,Y,X);
for i=2:numOfcomp
pmap(lista{i})=pmap(lista{i})+train_values(i).cp.CorrectRate;
spmap(lista{i})=spmap(lista{i})+train_values(i).cp.Sensitivity;
semap(lista{i})=semap(lista{i})+train_values(i).cp.Specificity;
counter(lista{i})=counter(lista{i})+1;
end
pmap=pmap./counter;
pmap=reshape(pmap,[Z Y X]);
spmap=spmap./counter;
spmap=reshape(spmap,[Z Y X]);
semap=semap./counter;
semap=reshape(semap,[Z Y X]);
%figure
%bmp_stack(pmap,5);
fprintf('Fin! \n')
%% Agregar votos
agregadmethod='mayoria'; %Puede introducirse como input
[prec]=agregar_votos(train_values,test_values,agregadmethod);
%%
res.TestValues=test_values;
res.TrainValues=train_values;
res.maps.pmap=pmap;
res.maps.spmap=spmap;
res.maps.semap=semap;
res.testset=test;
res.compList=lista;
res.Precision=prec;
end
function [lista]=listar(stack_all,jump,comp_size)
[~,masky]=mascara(stack_all);
Tama=size(stack_all);fprintf .
Z=Tama(2);Y=Tama(3);X=Tama(4);
grid= zeros(Z,Y,X);
for xx=1:jump:X
for yy=1:jump:Y
for zz=1:jump:Z
grid(zz,yy,xx)= 1;
end
end
end
lista_ind=find(grid);
ccomp=intersect(lista_ind,find(~masky));
lista=cell(1);
for r=1:numel(ccomp)
[z y x]=ind2sub([Z Y X],ccomp(r));
listap=[];
for m=1:comp_size
for n=1:comp_size
listap=[listap z+(y-2+n)*Z+(x-2+m)*Z*Y:z+(y-2+n)*Z+(x-2+m)*Z*Y+comp_size-1]; % min(,Z*Y*X)];
end
end
lista{r}=listap;
end
end
function [values]=component_accuracy(tr_data,tr_labels,nfolds,pname)
kernels= { 'linear' }; %'rbf' 'polynomial' 'quadratic' };
P=size(tr_data,1);
NK=numel(kernels);
values=struct('cp',cell(1,NK),'trad',cell(1,NK));
okopts={'svm','bayes'};
% pval = varargin{j+1};
cl = find(strncmpi(pname, okopts,numel(pname)));
switch cl
case 1
for kernel= 1:NK
%
cp=classperf(tr_labels); fprintf .
if nfolds==0;
indices=(1:P)';
nfolds=P;
else
indices = crossvalind('Kfold',P,nfolds); fprintf .
end
options=optimset('MaxIter',1000);
for p=1:nfolds
test = (indices == p); train = ~test;
train_data=tr_data(train,:);
test_data=tr_data(test,:);
svmStruct = svmtrain(double(train_data), tr_labels(train),'Kernel_Function',char(kernels(kernel)),'method','QP','options',options);
classes = svmclassify(svmStruct,test_data);
classperf(cp, classes,test,'Positive', 1, 'Negative', 0);
end
values(kernel).cp=cp; fprintf .
values(kernel).trad=svmStruct; fprintf .
fprintf('completado! \n')
end
case 2
cp=classperf(tr_labels); fprintf .
if nfolds==0;
indices=(1:P)';
nfolds=P;
else
indices = crossvalind('Kfold',P,nfolds); fprintf .
end
for p=1:nfolds
test = (indices == p); train = ~test;
train_data=tr_data(train,:);
tt_data=tr_data(test,:);
nb = NaiveBayes.fit(train_data, tr_labels(train));
class = predict(nb, tt_data);
classperf(cp, double(class),test,'Positive', 1, 'Negative', 0);
end
values.cp=cp; fprintf .
values.trad=nb; fprintf .
end
end
function [values]=component_minSV(tr_data,tr_labels,pname)
kernels= { 'linear' }; %'rbf' 'polynomial' 'quadratic' };
P=size(tr_data,1);
NK=numel(kernels);
values=struct('nSv',cell(1,NK),'trad',cell(1,NK));
okopts={'svm','bayes'};
% pval = varargin{j+1};
cl = find(strncmpi(pname, okopts,numel(pname)));
switch cl
case 1
for kernel= 1:NK
%
options=optimset('MaxIter',1000);
svmStruct = svmtrain(double(tr_data), tr_labels,'Kernel_Function',char(kernels(kernel)),'method','QP','options',options);
numofSV=numel(svmStruct.SupportVectorIndices);
values(kernel).trad=svmStruct; fprintf .
values(kernel).nSv=numofSV;
fprintf('completado! \n')
end
end
end