Spatial weight matrix is the popular numerical quantification of spatial dependency or spatial neighborhoods. The weight matrix should summarize information about the spatial connectivity structure of the spatial entities/features; either polygons, points, or lines. This is required for the computation of spatial dependency indices such the Moran’s index, and for spatial regression models such as the conditional autoregressive (CAR), spatial lag, and spatial error models. The connectivity information can be defined based on adjacency/contiguity or distance between pairs of spatial entities. There are other forms; they could be based on population densities between observation pairs. The simplest spatial weigh matrix is the binary adjacency spatial weight matrix with elements w_ij, such that w_ij=1 if spatial units i and j are neighbors, otherwise w_ij=0. A popular alternative is the inverse distance weight matrix with elements w_ij=1⁄d^α , where d is the distance between pairs of spatial units and α is any positive number greater than zero. By convention, w_ii=0 since spatial unit cannot have a spillover within itself.
Ability to define elements and requirements of the spatial weight matrix
Identify the different methods for constructing spatial weigh matrix
Compare the different types of spatial weight matrices
Completed (GI-N2K)