Predictive Moran's Eigenvector Maps

Calculate Predictive Moran's Eigenvector Maps (pMEM) for spatially-explicit prediction of environmental variables, as defined by Guénard and Legendre (2024) . pMEM extends classical MEM by enabling interpolation and prediction at unsampled locations using spatial weighting functions parameterized by range (and optionally shape). The package implements multiple pMEM types (e.g., exponential, Gaussian, linear) and features a modular architecture that allows programmers to define custom weighting functions. Designed for ecologists, geographers, and spatial analysts working with spatially-structured data.