Since you found a way to model a Lagrangian without an analytical solution, wouldn't it be interesting to throw data from systems we don't usually assume to be,

Since you found a way to model a Lagrangian without an analytical solution, wouldn't it be interesting to throw data from systems we don't usually assume to be, and find out if they could be modelled as one by looking at the error?