NEWS

GWPR.light 1.0.0 (2026-05-29)

New public API: the primary interface is now four snake_case functions: gwpr(), select_bandwidth(), fit_gwpr(), diagnose_gwpr().
sf first: spatial input must be an sf object. The legacy sp / SpatialPolygonsDataFrame interface of the 0.x API is not supported by the new functions.
Old functions deprecated: GWPR(), bw.GWPR(), GWPR.moran.test(), GWPR.pFtest(), GWPR.phtest(), and GWPR.plmtest() remain in the package for backwards compatibility but will be removed in a future version.

gwpr(): one-call pipeline that validates inputs, optionally searches for the optimal bandwidth, fits the model, and runs diagnostics.
select_bandwidth(): standalone bandwidth optimisation with three search strategies:
- "grid" — exhaustive grid search over a user-defined range.
- "sgd" — stochastic gradient descent (fast, scales to large grids).
- "random" — random sampling over a candidate set.
fit_gwpr(): fit a GWPR model with a known bandwidth. Supports family = "gaussian" (linear) and family = "binomial" (panel logistic).
diagnose_gwpr(): collects results from Moran's I ("moran"), local F-test ("f_test"), Hausman test ("hausman"), and LM test ("lm_test").
Five kernel functions: "bisquare" (default), "gaussian", "exponential", "tricube", "boxcar".
Adaptive bandwidth (adaptive = TRUE) using k-nearest-neighbour approach.
Optional parallelism via workers argument (requires future and future.apply; default workers = 1 is always safe).
Memory estimation: gwpr() warns when the estimated memory footprint is high before starting long computations.
Reproducibility: seed argument propagates to all stochastic steps.

Added: fixest, glmmTMB (for binomial panel logistic models).
future and future.apply moved to Suggests (only needed for workers > 1).
R (>= 3.5.0) minimum version (was R (>= 2.10)).