Calibration of Movement Models

Module for Movement Model calibration.

Every Movement Model is defined in terms of user controllable parameters which must be calibrated before use. Model calibration refers to the three following procedures:

  1. Velocity calibration:
    Even though all Movement Models must implement some way of controlling agent mean speed, it may be impossible to precisely select the correct speed parameter to obtain the desired mean speed, specially for complex models with interactions between environment and other individuals. Hence a correction to the speed parameter must be calculated to amend such shortcomings. Mean speed is calculated in a variety of scenarios and a correction factor is calculated by fitting the velocity data of the simulated movements to desired outcome.
  2. Home Range calibration.
  3. Occupancy calibration.
    A main feature of ollin is its capacity to generate random scenarios with a user determined home range and occupancy values. In order for this to be possible, the fundamental relations between home range and mean velocity, and density and occupancy, must be understood. These relations depend on the Movement Model. Home range and Occupancy calibration refer to the process in which simulations in a variety of scenarios are run and a parametric model of response functions are fitted to the data to obtain an approximate functional relation between home range and mean velocity, and density and occupancy.
ollin.calibration.calibrate.calibrate(model, config=None, save_fig=False, save_path=None, plot_style='fivethirtyeight')[source]

Calibrate parameters of movement model.

Parameters:
  • model (MovementModel) – Movement model instance to calibrate.
  • config (dict, optional) – Dictionary holding calibration configuration values. If None is given, default configurations will be used. See config
  • save_fig (bool, optional) – If True, calibration procedure will save calibration graphs.
  • save_path (str, optional) – Path in which to save calibration figures. Must be provided if save_fig is True.
  • plot_style (str, optional) – Name of pyplot style to use in calibration figures. See pyplot to see all available options.
Returns:

  • calibrated_model (MovementModel) – Same Movement model instance with calibrated parameters.
  • calibrated_parameters (dict) – Dictionary holding all calibrated parameters.
Raises:

ValueError: – If no save_path was provided while save_fig is True.