-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbi_Main.m
296 lines (257 loc) · 13.5 KB
/
bi_Main.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
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
%%%% To Do List: (1)Th = Topt + dT
%%%% (2)Catch the chol exceptions and by-pass such cases
! copy bi_functn.m bi_functn.m
Files = dir('./*.csv');
NumOfFiles = length(Files);
R1 = [0.25,0.5,0.75,0.95];
% R1 = 1;
% R2 = R1;
R = length(R1);
nopt = 11; %%% NUMBER OF PARAMETERS TO BE OPTIMISED!
MaxMargin = 30;
MindT = 1;
Transform.Aopt1 = 0; Transform.Aopt2 = 0;
Transform.aopt1 = 0; Transform.aopt2 = 0;
Transform.Ah1 = 0; Transfomr.Ah2 = 0;
Transform.ah1 = 0; Transform.ah2 = 0;
CASE = [0]; %%% 0: uniform; 1: log-transformed; 2: exp-transformed
CnF = [0.5];
for CN = 1:length(CnF)
cnF = CnF(CN);
for CC = 1:length(CASE) %%% disregard the last case, which we consider as not of great importance
Case = CASE(CC);
for Nf = 1:NumOfFiles %%%%
DATA = readtable(Files(Nf).name);
NumOfSamples = width(DATA)/2;
% WinL = size(DATA,1)*0.1;
% WinL = size(DATA(:,2*j-1:2*j),1)*0.1;
% WinL = 10;
Title = Files(Nf).name(1:end-4);
IDX = find(Title=='_');
IDX = IDX(1);
Acronym = [Title(1),upper(Title(IDX+1:IDX+2))];
% FdlName = ['./Result_WinL=length0.2_TMV',Acronym,'_Case',num2str(Case),'_cn',num2str(cnF*10)];
FdlName = ['Bi_Result_', Title, '_WinL=length0.1'];
% FdlName = ['Bi_Result_', Title, '_WinL=10','_cn', num2str(cnF*10)];
if ~exist(FdlName)
mkdir(FdlName);
end
ResultFileName = [Title,'_result','.csv'];
StatFileName = [Title,'_stat','.csv'];
X = ones(NumOfSamples,nopt+2)*nan;
RMS = ones(NumOfSamples,1);
ACCmean = zeros(NumOfSamples,1);
ACCmaxi = ACCmean;
ACCmini = ACCmaxi;
meanPAR = zeros(NumOfSamples,nopt);
stdPAR = meanPAR;
skewPAR = meanPAR;
kurtPAR = meanPAR;
for j = 1:NumOfSamples
LDATA = size(DATA(:,2*j-1:2*j));
WinL = size(DATA(:,2*j-1:2*j),1)*0.1;
if LDATA(1)>=3*WinL
x = ones(1,nopt)*nan;
close all
MAX = -1e+012;
data = table2array(DATA(:,2*j-1:2*j));
I = find(~isnan(data(:,1)));
data = data(I,:);
T = data(:,1);
obs = smoothdata(data(:,2),'movmean',WinL);
obs_raw = data(:,2);
% obs = data(:,2);
T = T+273.15;
[~,ia,~] = unique(T); T = T(ia); obs = obs(ia);
% plot(T,obs)
%%%% copy and rename your objective function as functn.m
xdata = T; ydata = obs;
Range = max(xdata) - min(xdata); %%%% Assume: the sample dies at the very end of the experiment so the feasible temperature range has been THOROUGHLY EXPLORED
for k1 = 1:R
r = R1(k1);
% for k2 = k1+1:R-1
% r2l = R1(k2);
% r2u = R1(k2+1);
disp(['Now computing ', Acronym, ' sample ' ,num2str(j),' CnFac = ',num2str(cnF),' Case ',num2str(Case) ' r = ', num2str(r)]);
% Define the lower/upper bounds of parameters of the objective function
%%% UNQUOTE TO GET 12 OR 14 PARAMETERS MODEL
if Case == 0 %%% baseline of x7: min(xdata)
bl= [log(0.1), (r-0.25)*Range, log(10), log(10), 0.01,...
log(0.1), r*Range, log(10), log(10), 0.01, 1e-6];
%% Topt2 = Topt1+dT
bu= [log(1000), r*Range, log(5000000), log(5000000), 30,...
log(1000), Range, log(5000000), log(5000000), 30, min(ydata)-1e-6];
elseif Case ==1 %%% base of x7: min(xdata)
bl= [log(0.1), log((r-0.25)*Range+0.0001), log(10), log(10), log(0.01),...
log(0.1), log(r*Range), log(10), log(10), log(0.01), 1e-6];
%% Topt2 = Topt1+dT
bu= [log(1000), log(r*Range), log(5000000), log(5000000), log(30),...
log(1000), log(Range), log(5000000), log(5000000), log(30), min(ydata)-1e-6];
elseif Case == 2 %%% base of x7: x2
bl= [log(0.1), log((r-0.25)*Range+0.0001), log(10), log(10), log(0.01),...
log(0.1), log(0.01), log(10), log(10), 0.01, 1e-6];
%% Topt2 = Topt1+dT
bu= [log(1000), log(r*Range), log(5000000), log(5000000), log(30),...
log(1000), log(MaxMargin), log(5000000), log(5000000), log(30), min(ydata)-1e-6];
end
s = 2000;
q = 20;
L = (s/q)/10;
T = 1e+06;
cn = 2.4/sqrt(nopt)*cnF;
M = 10000;
%cn = 0.4;
[bestf,bestx,NIter,D,Seq,acc,flag,diffpct,diffabs,GR,orig,new,SeqExc] = SCEM(nopt,s,q,bl,bu,xdata,ydata,Case,Transform,L,T,cn,M);
%%%% If SEM is terminated due to chol failure, then
%%%% bestf = -1e+012;
disp(['GR = ',num2str(GR)]);
MAXDIFF = max(abs(diffpct(:)));
MAXDIFF_abs = max(abs(diffabs(:)));
disp(['max(DIFF) = ',num2str(MAXDIFF),'%'])
disp(['max(DIFFabs) = ',num2str(MAXDIFF_abs)])
if bestf > MAX
MAX = bestf;
x = bestx;
SeqN = Seq;
NITER = NIter;
ACC = acc;
ACCmean(j) = mean(acc/NITER); %%% mean/max/mini across different sequences
disp(['mean acc rate:',num2str(ACCmean(j))]);
ACCmaxi(j) = max(acc/NITER);
disp(['max acc rate:',num2str(ACCmaxi(j))]);
ACCmini(j) = min(acc/NITER);
disp(['min acc rate:',num2str(ACCmini(j))]);
DiffPct = diffpct;
DiffAbs0 = diffabs;
Padding = zeros(1,nopt);
Padding(1) = MAXDIFF_abs;
DiffAbs = [Padding;DiffAbs0];
end
% end
end
PARAMS=[];
offset = 200;
for i = 1:q
PARAMS = [PARAMS;SeqN(10000*(q-1)+offset+1:10000*(q-1)+NITER,:)];
end
meanPAR(j,:) = nanmean(PARAMS(:,1:end-1));
stdPAR(j,:) = nanstd(PARAMS(:,1:end-1));
skewPAR(j,:) = skewness(PARAMS(:,1:end-1));
kurtPAR(j,:) = kurtosis(PARAMS(:,1:end-1));
if Case == 0
OrigPar2 = x(2);
OrigPar5 = x(5);
OrigPar7 = x(7);
X7_Baseline = min(xdata);
OrigPar10 = x(10);
OrigPar11 = x(11);
else
OrigPar2 = exp(x(2));
PARAMS(:,2) = exp(PARAMS(:,2));
OrigPar5 = exp(x(5));
PARAMS(:,5) = exp(PARAMS(:,5));
OrigPar7 = exp(x(7));
PARAMS(:,7) = exp(PARAMS(:,7));
if Case == 1
X7_Baseline = min(xdata);
else
X7_Baseline = min(xdata)+OrigPar2;
end
OrigPar10 = exp(x(10));
PARAMS(:,10) = exp(PARAMS(:,10));
OrigPar11 = x(11);
PARAMS(:,11) = PARAMS(:,11);
end
x1 = exp(x(1));
x2 = min(xdata)+OrigPar2;
x3 = exp(x(3));
x4 = exp(x(4));
x5 = x2+OrigPar5; %%% exp(x(5)): delta_T
x6 = exp(x(6));
x7 = X7_Baseline+OrigPar7;
x8 = exp(x(8));
x9 = exp(x(9));
x10 = x7+OrigPar10;
x11 = OrigPar11;
F1 = x1 * xdata/x2 .* exp(x3*(1/x2-1./xdata))./(1+exp(x4*(1/x5-1./xdata)));
F2 = x6 * xdata/x7 .* exp(x8*(1/x7-1./xdata))./(1+exp(x9*(1/x10-1./xdata)));
MDL = F1+F2+x11;
if x5 < x10
res = [x1,x6,x2-273,x7-273,x3,x8,x4,x9,x5-273,x10-273,x11,GR,MAXDIFF_abs];
else
res = [x6,x1,x7-273,x2-273,x8,x3,x9,x4,x10-273,x5-273,x11,GR,MAXDIFF_abs];
end
X(j,:) = res;
% X_raw(j,:) = [x1,x6,x2-273,x7-273,x3,x8,x4,x9,x5-273,x10-273,flag];
RMS(j) = rms(MDL-obs);
close all;
figure(1);
histogram(min(xdata)+PARAMS(:,2)-273,20,'Normalization','probability'); hold on;
xline(x2-273,'--b','linewidth',1.4); hold on;
histogram(X7_Baseline+PARAMS(:,7)-273,20,'Normalization','probability'); hold on;
xline(x7-273,'--r','linewidth',1.4);
xlabel('Topt');
title([Title,'Topt Case',num2str(Case),',cn',num2str(10*cnF)]);
FileName = [FdlName,'/',Title,'_',num2str(j),'Topt'];
saveas(gcf, FileName, 'pdf');
figure(2);
histogram(min(xdata)+PARAMS(:,2)+PARAMS(:,5)-273,20,'Normalization','probability'); hold on;
xline(x5-273,'--b','linewidth',1.4);
histogram(X7_Baseline+PARAMS(:,7)+PARAMS(:,10)-273,20,'Normalization','probability'); hold on;
xline(x10-273,'--r','linewidth',1.4);
xlabel('T_H');
title([Title,'TH Case',num2str(Case),',cn',num2str(10*cnF)]);
FileName = [FdlName,'/',Title,'_',num2str(j),'T_H'];
saveas(gcf, FileName, 'pdf');
figure(3);
h1=plot(xdata-273,MDL-x11); hold on; h2=plot(xdata-273,obs-x11);hold on;
plot(xdata-273,F1,'k--'); hold on; plot(xdata-273,F2,'m--'); hold on;
plot(data(:,1),obs_raw-x11,'r--','linewidth',1.4);
xlabel('Temp (Celsius)');
ylabel('HR (bp/min.)');
title([Title,'Rslt Case',num2str(Case),',cn',num2str(10*cnF)]);
legend([h1,h2],{'Model','Obs.'},'location','best');
FileName = [FdlName,'/',Title,'_',num2str(j),'_MDL'];
saveas(gcf, FileName, 'png');
if max(abs(DiffPct(:)))>0
DIFFpctName = [Title,'_',num2str(j),'_DiffPct','.csv'];
writematrix(DiffPct,[FdlName,'/',DIFFpctName]);
end
if max(abs(DiffAbs(:)))>0
DIFFabsName = [Title,'_',num2str(j),'_DiffAbs','.csv'];
writematrix(DiffAbs,[FdlName,'/',DIFFabsName]);
end
end
end
% RESULTS = array2table([X,RMS,ACCmean,ACCmaxi,ACCmini],'VariableNames',...
RESULTS = array2table([X(:,1:10),RMS,X(:,11:13),ACCmean,ACCmaxi,ACCmini],'VariableNames',...
{'Rho_ref_1', 'Rho_ref_2',...
'T_ref_1', 'T_ref_2',...
'dHa_R_1', 'dHa_R_2',...
'dH_high_R_1', 'dH_high_R_2',...
'T_high_1', 'T_high_2',...
'rms',...
'Bp0','GR','MaxAbsErr',...
'ACCmean','ACCmaxi','ACCmini'});
writetable(RESULTS,[FdlName,'/',ResultFileName]);
STATS = array2table([meanPAR(:,2),stdPAR(:,2),skewPAR(:,2),kurtPAR(:,2),...
meanPAR(:,5),stdPAR(:,5),skewPAR(:,5),kurtPAR(:,5),...
meanPAR(:,7),stdPAR(:,7),skewPAR(:,7),kurtPAR(:,7),...
meanPAR(:,10),stdPAR(:,10),skewPAR(:,10),kurtPAR(:,10)],'VariableNames',...
{'meanP2','stdP2','skewP2','kurtP2',...
'meanP5','stdP5','skewP5','kurtP5',...
'meanP7','stdP7','skewP7','kurtP7',...
'meanP10','stdP10','skewP10','kurtP10'});
writetable(STATS,[FdlName,'/',StatFileName]);
% RESULTS_raw = array2table([X_raw,RMS],'VariableNames',...
% {'Rho_ref_1', 'Rho_ref_2',...
% 'T_ref_1', 'T_ref_2',...
% 'dHa_R_1', 'dHa_R_2',...
% 'dH_high_R_1', 'dH_high_R_2',...
% 'T_high_1', 'T_high_2',...
% 'flag', 'rms'});
% writetable(RESULTS_raw,['./Results2/',ResultFileName_raw]);
%
end
end
end