I personally prefer to have the modelled value on the x-axis, but that is because I am often dealing with physics experiments where the modelled variable is a well controlled variable and the observed variable contains the error. Plotting the modelled variable on the x-axis can also be a way to decrease the dimensionality.
If you have multiple variables that you control in an experiment then it can be difficult to plot the output/observed value as function of all those variables (because it is multidimensional). But instead of all those variables you can replace them by the modelled variable. This might be your case as you could view your data as the 3d plot below
I personally prefer to have the modelled value on the x-axis, but that is because I am often dealing with physics experiments where the modelled variable is a well controlled variable and the observed variable contains the error. Plotting the modelled variable on the x-axis can also be a way to decrease the dimensionality. If you have multiple variables that you control in an experiment then it can be difficult to plot the output/observed value as function of all those variables (because it is multidimensional). But instead of all those variables you can replace them by the modelled variable.
I personally prefer to have the modelled value on the x-axis, but that is because I am often dealing with physics experiments where the modelled variable is a well controlled variable and the observed variable contains the error. Plotting the modelled variable on the x-axis can also be a way to decrease the dimensionality.
If you have multiple variables that you control in an experiment then it can be difficult to plot the output/observed value as function of all those variables (because it is multidimensional). But instead of all those variables you can replace them by the modelled variable. This might be your case as you could view your data as the 3d plot below