Project intends to fit Kuramoto to provided data. Currently there are two available scenarios:
- all-to-all, where fit is for N separated oscillators, and
- all-to-one, where fit is for signal that implicitly is composed of N oscillators.
In all cases the model is ran from respective directory with
$ python kuramoto.py
If this is run for the first time it might take a while to execute as Stan needs to compile the model. Another script is provided for displaying and comparing reconstruction:
$ python plot.py
Parameters for the models are stored in model.config
. Currently this is used for generating model for tests and as initial values.
Input is supposed to be of dimension NxT, i.e. N oscillators and their T phases in time or YN(t) = (y1(t), y2(t), ..., yN(t)). The model fits initial positions, couplings and intrinsic frequencies.
The input is a linear superposition of many oscillators: SN(T) = s1(t)+s2(t)+...+sN(t). The model fits initial positions, couplings and intrinsic frequencies.