Constructing Adjacency Matrices Based on Spatial and Feature Similarity // Constructs sparse adjacency matrices from spatial coordinates, feature measurements, class labels, and temporal indices. Supports nearest-neighbor graphs, heat-kernel weights, graph Laplacians, diffusion operators, and bilateral smoothers for graph-based data analysis, following spectral graph methods in von Luxburg (2007) doi:10.1007/s11222-007-9033-z, diffusion maps in Coifman and Lafon (2006) doi:10.1016/j.acha.2006.04.006, and bilateral filtering in Tomasi and Manduchi (1998) doi:10.1109/ICCV.1998.710815.