Simulate and calibrate linear propagator models for price responses to an external order flow. The models and methods are explained and applied to real high-frequency trading data in:
Patzelt, F. and Bouchaud, J-P. (2017): Nonlinear price impact from linear models. Journal of Statistical Mechanics: Theory and Experiment, 12, 123404. Preprint at arXiv:1708.02411.
| Function | Synopsis |
|---|---|
| G_pow | Return power law Propagator kernel |
| beta_from_gamma | Return exponent beta for a power law propagator kernel that decorrelates an input with a pure power law autocorrelation with exponent gamma |
| calibrate_hdim2 | Calibrate two-kernel History Dependent Impact Model |
| calibrate_tim1 | Calibrate original Transient Impact Model |
| calibrate_tim2 | Calibrate two-kernel Transient Impact Model |
| hdim2 | Simulate two-kernel History Dependent Impact Model |
| integrate | Return lag 1 sum, i.e. convert a differential kernel to a "bare response". |
| k_pow | Return differential form of power law propagator kernel |
| propagate | Apply propagator kernel to a time series (FFT conv.) |
| response | Calculate e.g. a price response |
| response_grouped_df | Calculate response for pandas groups and average |
| smooth_tail_rbf | Smooth the tail of a long kernel using logarithmically spaced Radial Basis Functions |
| tim1 | Simulate original Transient Impact Model |
| tim2 | Simulate two-kernel Transient Impact Model |
The submodule batch automates model calibration and simulation. Please
find further explanations in the docstrings and in the examples directory.
The required methods to efficiently estimate two- and three-point correlation matrices were released in the separate package scorr.
pip install priceprop
- Python 2.7
- NumPy
- SciPy
- Pandas
- scorr
- Jupyter
- Matplotlib
- colorednoise