I am working with Phoenix models using the new version of the program: I am find that the models I have accumulated are now working fine. I have one feature that I do not understand. In the attached photo I have marked a rectangle of interest. When performing a simulation it is convenient to have a series of specified integer time concentrations with corresponding plasma concentrations. Therefore after being successful with the running the model I became more ambitious and attempted to get the X, Y coordinates simulated specified above.
I was not really sure how to do it so I wonder if I could have some advice on the steps. The zone of interest is marked with a rectangle {I am aware that C should be C1).
On the right of times, you write the times at which you want the responses to be simulated.
There is a short cut you can used which is the seq statement.
seq(0,10,1)
means simulate times starting at 0 up to 10 by intervals of 1.
It is equivalent to write on the right of times
0,1,2,3,4,5,6,7,8,9,10
Now on the right of variables, you write the responses of interest like C but you can add also any additional variables that is defined like covariate, observation like CObs.
CObs should simulate around C from the error distribution and would give you what could be a real observation if the model was correct but you have noise in your observed data reflected by the error model.
Serge: Thank you it does help. It works fine! One thing is not clear from your remarks is as follows: my model does not contain values for the experimentally observed plasma concentrations. It is a simulation model. it contains values for the Ka and Ke and V at each of the 5 doses at the 5 different times. The Ka is different at Aa1 compared with the other 4 doses. The other 4 doses have a lower value for Ka. My model simulates values of the plasma concentrations and I am calling that Cobs as shown in my model diagram. Do you find this confusing? Or am I using the program incorrectly.
The Cobs term is calculated by simulation from the model; actually it should be called Cpre. I wonder if I can change that to Cpre for predicted in my model.
Additionally I am not clear regarding the essential difference between C and Cobs that I get in my output. What is definition C.?
I do have the actual experiment plasma concentrations from the lab.at study time points. They are the actual Cobs concentrations. My plan was to perform a SIMPLE (FITTING) using these plasma concentrations with parameter estimates, but I am not sure if that will work (too may variables).
C is the predicted response. CObs is the predicted response + noise. This represents what you could have gotten in a new experiment other than the one you observe.
