Hey Matlab friends, I have to estimate my 4 parameters, for 200 .matfiles and store the results. I created the codes below but it works for the first 6 datasets.

files = dir('*.mat') ; % all the .mat files in the folder

N = length(files) ; % Total number of files

for LL=1:N

thisfile = files(LL).name ;

load(thisfile)

DV=T.dv;

lengthTable=length(DV);

ID=T.ID;

time=T.time;

dose=T.amt;

DVnew=[];

timeNew=[];

doseNew=[];

IDnew=[];

lengthTable=length(DV);

k=1;j=1;

for i=1:lengthTable

if ~isnan(dose(i))

doseperid(k)=dose(i);

k=k+1;

else

if ~isnan(time(i))&&(~isnan(DV(i)))&&(~isnan(ID(i)))

doseNew(j,:)=doseperid(k-1);

timeNew(j,:)=time(i);

DVnew(j,:)=DV(i);

IDnew(j,:)=ID(i);

j=j+1;

end

end

end

dosePerID=doseperid';

%% SSEMMpk with ode45

logparameter0=log([1 1 0.2 0.5 ]);

model = @Getsolutionode;

n=length(dosePerID);

%% SSEMMpk with ode45

P=[1 0 0 0 ; 0 1 0 0; 0 0 1 0; 0 0 0 1];

%% nlmefit

[beta,PSI,stats,b] = nlmefit(timeNew,DVnew,IDnew,...

dosePerID,model,logparameter0,...

'REParamsSelect',[1 2 3 4 ],'ErrorModel','Constant','ApproximationType','FOCE','CovPattern',P,'RefineBeta0','on',...

'ParamTransform',repelem(1,4),'OptimFun','fminunc');

%% nlmefir output

beta2=beta;

b2=b;

bb=exp(beta);

psi=PSI;

STATS=stats;

STATS.COVB=stats.covb;

parameter=exp(beta');

predictor=timeNew;

%% storing all the dataset outputs

Beta{:,LL}=beta2;

B{LL}=b2;

A2{:,LL}=bb;

psi2{:,LL}=psi;

STATS2{:,LL}=STATS;

STATS.COVB2{:,LL}=STATS.COVB;

end

function concentration=Getsolutionode(parameter,t,doseNew)

%% initial values

ka=parameter(:,1);

v=parameter(:,2);

vmax=parameter(:,3);

km=parameter(:,4);

predictor(:,1)=t;

minTime=0;

maxTime=max(t);

y0=[doseNew 0];

%% ode45

options=odeset('RelTol',1e-6);

solode=ode45(@odefunction,[minTime maxTime], [y0],options,ka,v,vmax,km);

rspl=deval(solode,t);

concentration=rspl(2,:)'./v;

end

[Stephen23]

Do not LOAD directly into the workspace, always LOAD into an output variable (which is a scalar structure):

S = load(..);

and then access its fields. This make you code more robust and might help you debug this problem too.

[Sepi Sharif]

Thanks Stephen for your comment. My problem should be something about storing or looping this nlmefit through 200 datasets? which I cant find t out. it is going and storing for the first six datasets but not more!

[Sepi Sharif]

it might be helpful, so when I run each datset separately, it works, gave me the estimations and all information, but through the loop for all of them and storing ithe output, it faced with error after 6 loops!

[Stephen23]

" it faced with error after 6 loops!"

Please show us the complete error message. This means all of the red text.

[Sepi Sharif]

update:

I have changed the first two lines of loading

files = dir('*.mat') ; % all the .mat files in the folder

N = length(files) ; % Total number of files

to:

files = dir('*.mat') ; % all the text files in the folder

N = length(files) ;

the error :

Error using nlmefit>checkFunVals (line 1858)

MYFUN has returned Inf or NaN values.

Error in nlmefit>GNfitPNLS (line 1267)

if funValCheck && ~isfinite(sse_new), checkFunVals(sse_new); end

Error in nlmefit (line 601)

[BetaNew,bNew,stop] = GNfitPNLS(Delta,Y,Beta,b,weights,PNLSOpt,grp,grpMatch,M,N,r,q,computeFAllGroupsFcn,computeJacobianAllGroupsFcn);

Error in MMpk_IndLin_Test (line 80)

[beta,PSI,stats,b] = nlmefit(timeNew,DVnew,IDnew,...