pMEM: Predictive Moran's Eigenvector Maps for Spatial Modeling
pMEM implements Predictive Moran's Eigenvector Maps, a method for
spatially-explicit prediction of environmental variables using
eigen-decomposition of distance-based spatial weighting matrices.
Guénard, G., & Legendre, P. (2024). Spatially-explicit predictions using spatial
eigenvector maps. Methods in Ecology and Evolution.
https://doi.org/10.1111/2041-210X.14413
What Does pMEM Do?
pMEM extends classical Moran's Eigenvector Maps (MEM) by:
-
Enabling prediction at unsampled locations via
predict() methods
-
Supporting tunable distance weighting functions (exponential, Gaussian,
power, linear, spherical, hyperbolic, hole_effect)
-
Optimizing spatial scale parameters via cross-validated model selection
-
ntegrating with regression frameworks: linear models, GLMs, elastic net
-
Providing fast C++ backend via Rcpp for efficient eigenvector selection
-
Supporting asymmetric (directional) spatial processes via complex-valued
distance metrics
Ideal for ecologists, geographers, and spatial analysts modeling
spatially-autocorrelated variables such as:
- Environmental gradients (depth, temperature, substrate)
- Species distributions and abundances
- Landscape connectivity metrics
- Directional processes (river flow, wind patterns, ocean currents)
Installation
From CRAN (stable release)
install.packages("pMEM")