Estimation Module

Module defining basic estimation interface and library tools.

Passive detector data can be used to estimate ecological state variables. This module holds libraries of models used for estimation of these state variables.

Currently only occupancy estimation models are available, but we expect to add more state variables in the near future.

Notes

All estimation models must inherit from EstimationModel metaclass. Hence any estimation model must implement its EstimationModel.estimate() method. Estimation models must return an Estimate object. Estimate objects should specialize to whatever is being estimated, see OccupancyEstimate for an example.

New estimation models must be placed within their corresponding directories. For example, any new occupancy estimation model should be placed in:

ollin/estimation/occupancy/

Libraries of estimation models for other state variables can also be built following the same structure, say:

ollin/estimation/abundance/

could contain abundance estimation models.

Every estimation model must be contained in a new file, in its corresponding directory, so that get_estimation_model() and get_estimation_model_list() tools function correctly. These tools will look for files in these directories automatically and no further integration is needed.

class ollin.estimation.estimation.Estimate(model, data)

Basic estimate class.

Any estimate result must hold, apart from the estimate, the original data and the estimation model used.

This class is meant to serve as an interface for specialized estimates of different state variables. Occupancy, abundance and other state variables must have specialized Estimate classes.

model

EstimationModel – Model used for estimation.

data

Detection – Detection data used for estimation.

class ollin.estimation.estimation.EstimationModel

Basic estimation model class.

Any estimation model must subclass this class and define an estimate method.

name

str – Name of estimation model.

estimate(detection)

Make estimate using detection data.

Abstract method that must be overwritten by any estimation model. Its argument must be a Detection object, and it must return a Estimate.