Fitting function to ecdf plot

Hi

I am looking for a way to fit function g to the ecdf-plot I produce as follows.

T = 1000;
cens = (y>T);
ecdf(y,'Function','survivor', ...
    'Censoring',cens,'Bounds','on');    
g=fittype('a-b*exp(-c*x)');

I would like to use the bounds determined in ecdf as a weighting factor "w" for the fit.

I tried

h=gcf;
 dataObjs = findobj(gcf,'-property','YData');
 dmy3=dataObjs.XData;
 x=dmy3';
 dmy4=dataObjs.YData;
 y2=dmy4';
 g=fittype('a-b*exp(-c*x)');
 f=fit(x,y2,g,'StartPoint',[[ones(size(x)), -exp(-x)]\y2; 1]);
 hold on; 

Though this is giving me the upper bounds as xData and yData.

Thanks in advance for any suggestion.

Peter

Answer 1

[Peter Mühlenbrock]

Iw ill close this, as I found a solution.

%----access all xy data-----
d = findall(gcf, '-property', 'xdata');
xydatas = arrayfun(@(h) get(h, {'xdata','ydata', 'type'}), d, 'Uniform', 0);
%----adress xy data-----
x1=xydatas{3, 1}{1, 1}'  ;
y1=xydatas{3, 1}{1, 2}'  ;

w=xydatas{3, 1}{1, 2}'-xydatas{2, 1}{1, 2}';
 g=fittype('a*exp(-b*x)+c');
 f=fit(x1,y1,g,'StartPoint',[1, 0.02, 0.5],'Weight',w);
 hold on; 
plot(f)