To explore the identification of chaotic dynamics evolving on a finite dimensional attractor, let's consider the nonlinear Lorenz system:
The trapezoidal rule is a numerical method to solve ordinary differential equations that approximates solutions to initial value problems of the form:
The trapezoidal rule states that: The model will be trained so that the trapezoidal rule is satisfied. LHS will be the target and the RHS will be the sum of the outputs from the model multiplied by a time step.Following are the commands used to train and test the model:
To train the model:
$ python trapezoidal.py train --itr 2000
To test the model:
$ python trapezoidal.py test
Following is the obtained result: