how to use nlinfit

I have a data Y like this:

      0
    0.1386
    0.1875
    0.2315
    0.2315
    0.2125
    0.2062
    0.2125
    0.2062
    0.2572
    0.2188
    0.2572
    0.2315
    0.3164
    1.0670
    1.7491
    1.6116
    1.1873
    0.7250
    0.5233
    0.3641
    0.3031
    0.3097
    0.3164
    0.3367
    0.4131
    0.3435
    0.3367
    0.3164
    0.3367
    0.2898
    0.2636
    0.3299
    0.3031
    0.2767
    0.3164
    0.2898
    0.3367
    0.3232
    0.3031

Data X:

        0
      1.4300
      2.8600
      4.2900
      5.7200
      7.1500
      8.5800
     10.0100
     11.4400
     12.8700
     14.3000
     15.7300
     17.1600
     18.5900
     20.0200
     21.4500
     22.8800
     24.3100
     25.7400
     27.1700
     28.6000
     30.0300
     31.4600
     32.8900
     34.3200
     35.7500
     37.1800
     38.6100
     40.0400
     41.4700
     42.9000
     44.3300
     45.7600
     47.1900
     48.6200
     50.0500
     51.4800
     52.9100
     54.3400
     55.7700

This is a curve, which I need to approximate by a function

modelfun = @(k,t)(((1.*k(1))./(k(1)-k(2))).*(-exp(-k(1).*t)+exp(-k(2).*t)));

So, I need to estimate the parameters k(1) and k(2). Here I'm using nlinfit function

beta = nlinfit(X,Y,modelfun,beta0)

with

   beta0 = [0.1 0.2].

When I get the parameters, I want to get the opposite - generate data similar to Y, but when I do this by

Y = modelfun(beta,X)

I get the wrong results and I do not know why?????

[Star Strider]

You seem to be doing everything correctly. Plotting your data, it is obvious that your modelfun does not describe it.

Y has a peak at X = 23.45. Either you need to revise your model to accommodate the peak, or if the peak is noise, collect new data and fit your function to it.

Only you can determine what is most appropriate.