compute_derivatives: Compute numerical derivatives from data using weighted linear...

In PhilChodrow/compositional-complexity: Analyze the spatial complexity of compositional data

Description

Compute numerical derivatives from data using weighted linear regression on each tract's 1-ego network.

Usage

compute_derivatives(adj, r_sigma = 1, s_sigma = 1, smooth = F)

Arguments

adj	the adjacency matrix on which to compute. Must include centroids.
r_sigma	geographic bandwidth of the regression weight decay
s_sigma	if smooth = TRUE, the bandwidth of a geographic rbf smoother
smooth	boolean should the derivatives be computed on a geographically smoothed demographic distribution?

Value

a tract-level tibble with a column D_alpha giving the derivatives of demographic frequencies with respect to spatial coordinates.

PhilChodrow/compositional-complexity documentation built on May 8, 2019, 1:33 a.m.