function [risk,prediction]=physionet2012(time,param,value)
% PCA based
% [risk,prediction]=physionet2012(time,param,value)
%
% Sample Submission for the PhysioNet 2012 Challenge. Variables are:
%
% time - (Nx1 Cell Array) Cell array containing time of measurement
% param - (Nx1 Cell Array) Cell array containing type (param) of
% measurement
% value - (Nx1 Double Array) Double array containing value of measurement
%
%
% risk - (Scalar) estimate of the risk of the patient dying in hospital
% prediction - (Logical)Binary classification if the patient is going to die
% in the hospital (1 - Died, 0 - Survived)
%
% Example:
% [risk,prediction]=physionet2012(time,param,value)
%
% Written by Ikaro Silva, 2012
%
% Version 1.1
%
%The sample submission calculates the in hospital death
%probability based on Bayesian Rule conditioned on SAPS_SCORE
%and PDF fitting based on training Set A.
%
% P(dying | SAPS_SCORE) = P(dying)*P(SAPS_SCORE|Dying)/
% (P(dying)*P(SAPS_SCORE|dying) + P(living)*P(SAPS_SCORE|living))
%
%Calculate likelihood, P(SAPS_SCORE|Dying), based polynomial fitting of the CDF
%of the training data A conditioned on those that died (values of the fit
%are hardcoded below)
%
% saps_died=SAPS_SCORE(ihd==1); %ihd is logical vector where 1 = in hospital death
% MX_SAPS=4*14;
% [Ndied,xx]=hist(saps_died,[0:MX_SAPS]);
% md=nanmean(saps_died);
% st_d=nanstd(saps_died)
% pdf_died=normpdf(xx,md,st_d);
% plot(xx,Ndied./sum(Ndied));hold on;grid on;plot(xx,pdf_died,'r') %Check fit
%
% Repeat to get conditional probability on those that lived using Extreme
% Value Distribution
% saps_alive=SAPS_SCORE(ihd==0); %ihd is logical vector where 1 = in hospital death
% [Nalive,xx]=hist(saps_alive,[0:MX_SAPS]);
% parmhat = evfit(saps_alive(~isnan(saps_alive)));
% pdf_alive=evpdf(xx,parmhat(1),parmhat(2));
% figure
% %This fitting is not very good, but for the purpose of an example will do
% plot(xx,Nalive./sum(Nalive));hold on;grid on;plot(xx,pdf_alive,'r')
%
% TH=0.2700; %Threshold for classifying the patient as non-survivor
% % %Determined through maximization of the min(PPV,Sensitivity) on
% % %training Set A
% %
% TH_SAPS=.3300;
%
%
%
%
% B=[ 1.4037238316e+00;
% -8.1833389095e-02;
% -2.4939330155e-02;
% -1.5119424507e-02;
% 2.7833454606e-02;
% -1.4908319274e-02;
% 1.2317990562e-02;
% 1.0654354558e-02;
% 5.6335820205e-02;
% 2.0570191510e-01;
% -2.7707893188e-01;
% 4.0909316076e-04;
% -1.7088353456e-04;
% -5.4632240876e-03;
% 5.3569006748e-03;
% 1.2035782695e-02;
% -4.1221677189e-02];
%
%
% B_saps=[
% 4.8341800475e+00;
% 2.1514309096e-01;
% -2.3846361213e-01;
% ];
I=1;
i=1;
[ALL_CATEGORIES,time_series_names,descriptors]=get_param_names();
num_params=length(ALL_CATEGORIES);
num_ts_params=length(time_series_names);
num_descriptors=length(descriptors);
MEAN_DATA_24=zeros(I,num_ts_params) + NaN;
MEAN_DATA_48=zeros(I,num_ts_params) + NaN;
DESCRIPTORS=zeros(I,num_descriptors) + NaN;
[times,values,names]=extract_param_series(time,param,value);
[ts_times,ts_values,ts_names]=get_param_subset(time_series_names,times,values,names);
[des_times,des_values,des_names]=get_param_subset(descriptors,times,values,names);
DESCRIPTORS(i,:)=cell2mat(des_values(:))';
means24=calculate_mean(ts_times,ts_values,ts_names,time_series_names,[0 24*60]);
means48=calculate_mean(ts_times,ts_values,ts_names,time_series_names,[24*60 48*60]);
MEAN_DATA_24(i,:)=means24;
MEAN_DATA_48(i,:)=means48;
%SAPS_SCORES(i)=saps_score(time,param,value,1,[0 24]);
%SAPS_SCORES_48(i)=saps_score(time,param,value,1,[24 48]);
[PHAT,TH]=PCA_classify(MEAN_DATA_24,MEAN_DATA_48,DESCRIPTORS,time_series_names,des_names);
if isnan(PHAT(:,2))
X_trunc_saps=[SAPS_SCORES-SAPS_SCORES_48 SAPS_SCORES ];
PHAT_saps = mnrval(B_saps,X_trunc_saps);
end
if isnan(PHAT(2))
% [risk,prediction]=physionet2012_SAPS(time,param,value);
risk=0.5;
prediction=risk>TH;
elseif PHAT(2)<0.01
risk=0.01;
prediction=risk>TH;
elseif PHAT(2)>0.99
risk=0.99;
prediction=risk>TH;
else
risk=PHAT(2);
prediction=risk>TH;
end
function [PHAT,best_th]=PCA_classify(X1,X2,DESCRIPTORS,names1,names2)
% X_orig=X;
% x1=X(~isnan(X(:,1)) & ~isnan(X(:,2)),1);
% x2=X(~isnan(X(:,1)) & ~isnan(X(:,2)),2);
% X=[x1 x2];
% %n2=length(x1);
% DIF_A=X1-X2;
% DIF_A(:,strcmp(names1,'PH'))=abs(DIF_A(:,strcmp(names1,'PH')));
DIF_A=X2;
X=[X1 DIF_A ];
X=[DESCRIPTORS(:,strcmp(names2,'Age')) X];
ICUTYPE_a=[DESCRIPTORS(:,strcmp(names2,'ICUType'))];
if ICUTYPE_a==1
%savefile='icu1.mat';
savefile='icu_1.mat'; % for all
% savefile='icu_1_pca.mat';
% savefile='icu_all.mat';
load(savefile,'N2','PHAT_all','best_th','A','S','Mu','V', 'CV', 'HP', 'LC','N','pc1','mu','sigma','B');
S_new=calc_S_new_data(X',A,V,Mu,N,CV);
X_new_rec =( repmat(Mu,1,1) + A*S_new)';
sigma0 = sigma;
sigma0(sigma0==0) = 1;
z = bsxfun(@minus,X_new_rec', mu');
z = bsxfun(@rdivide, z, sigma0');
x_in=pinv(pc1)*z;
x_in=x_in(1:N2);
PHAT = mnrval(B,x_in');
elseif ICUTYPE_a==2
% savefile='icu2.mat';
savefile='icu_2.mat';
% savefile='icu_2_pca.mat';
% savefile='icu_all.mat';
load(savefile,'N2','best_th','A','S','Mu','V', 'CV', 'HP', 'LC','N','pc1','mu','sigma','B');
S_new=calc_S_new_data(X',A,V,Mu,N,CV);
X_new_rec =( repmat(Mu,1,1) + A*S_new)';
sigma0 = sigma;
sigma0(sigma0==0) = 1;
z = bsxfun(@minus,X_new_rec', mu');
z = bsxfun(@rdivide, z, sigma0');
x_in=pinv(pc1)*z;
x_in=x_in(1:N2);
PHAT = mnrval(B,x_in');
elseif ICUTYPE_a==3
% savefile='icu3.mat';
savefile='icu_3.mat';
% savefile='icu_3_pca.mat';
% savefile='icu_all.mat';
load(savefile,'N2','best_th','A','S','Mu','V', 'CV', 'HP', 'LC','N','pc1','mu','sigma','B');
S_new=calc_S_new_data(X',A,V,Mu,N,CV);
X_new_rec =( repmat(Mu,1,1) + A*S_new)';
sigma0 = sigma;
sigma0(sigma0==0) = 1;
z = bsxfun(@minus,X_new_rec', mu');
z = bsxfun(@rdivide, z, sigma0');
x_in=pinv(pc1)*z;
x_in=x_in(1:N2);
PHAT = mnrval(B,x_in');
elseif ICUTYPE_a==4
% savefile='icu4.mat';
savefile='icu_4.mat';
% savefile='icu_4_pca.mat';
% savefile='icu_all.mat';
load(savefile,'N2','best_th','A','S','Mu','V', 'CV', 'HP', 'LC','N','pc1','mu','sigma','B');
S_new=calc_S_new_data(X',A,V,Mu,N,CV);
X_new_rec =( repmat(Mu,1,1) + A*S_new)';
sigma0 = sigma;
sigma0(sigma0==0) = 1;
z = bsxfun(@minus,X_new_rec', mu');
z = bsxfun(@rdivide, z, sigma0');
x_in=pinv(pc1)*z;
x_in=x_in(1:N2);
PHAT = mnrval(B,x_in');
else
error('no icu type')
end
end
end