gwlearn.ensemble.GWRandomForestClassifier

class gwlearn.ensemble.GWRandomForestClassifier(*, bandwidth, fixed=False, kernel='bisquare', n_jobs=-1, fit_global_model=True, measure_performance=True, strict=False, keep_models=False, temp_folder=None, batch_size=None, min_proportion=0.2, undersample=False, random_state=None, verbose=False, **kwargs)

Generic geographically weighted random forest classifier

NOTE: local models leave out focal, unlike in traditional approaches. This allows assessment of geographically weighted metrics on unseen data without a need for train/test split, hence providing value for all samples. This is needed for futher spatial analysis of the model performance (and generalises to models that do not support OOB scoring).

Parameters:

bandwidthint | float

bandwidth value consisting of either a distance or N nearest neighbors

fixedbool, optional

True for distance based bandwidth and False for adaptive (nearest neighbor) bandwidth, by default False

kernelstr | Callable, optional

type of kernel function used to weight observations, by default “bisquare”

n_jobsint, optional

The number of jobs to run in parallel. -1 means using all processors by default -1

fit_global_modelbool, optional

Determines if the global baseline model shall be fitted alognside the geographically weighted, by default True

measure_performancebool, optional

Calculate performance metrics for the model, by default True

strictbool | None, optional

Do not fit any models if at least one neighborhood has invariant y, by default False. None is treated as False but provides a warning if there are invariant models.

keep_modelsbool | str | Path, optional

Keep all local models (required for prediction), by default False. Note that for some models, like random forests, the objects can be large. If string or Path is provided, the local models are not held in memory but serialized to the disk from which they are loaded in prediction.

temp_folderstr | None, optional

Folder to be used by the pool for memmapping large arrays for sharing memory with worker processes, e.g., /tmp. Passed to joblib.Parallel, by default None

batch_sizeint | None, optional

Number of models to process in each batch. Specify batch_size fi your models do not fit into memory. By default None

min_proportionfloat, optional

Minimum proportion of minority class for a model to be fitted, by default 0.2

undersamplebool, optional

Whether to apply random undersampling to balance classes, by default False

random_stateint | None, optional

Random seed for reproducibility, by default None

verbosebool, optional

Whether to print progress information, by default False

**kwargs

Additional keyword arguments passed to model initialisation

__init__(*, bandwidth, fixed=False, kernel='bisquare', n_jobs=-1, fit_global_model=True, measure_performance=True, strict=False, keep_models=False, temp_folder=None, batch_size=None, min_proportion=0.2, undersample=False, random_state=None, verbose=False, **kwargs)

Methods

__init__(*, bandwidth[, fixed, kernel, ...])
fit(X, y, geometry)	Fit the geographically weighted model
predict(X, geometry)
predict_proba(X, geometry)	Predict probabiliies using the ensemble of local models

fit(X, y, geometry)

Fit the geographically weighted model

Parameters:

Xpd.DataFrame

Independent variables

ypd.Series

Dependent variable

geometrygpd.GeoSeries

Geographic location