# Null deviance: 2838. # (Dispersion parameter for binomial family taken to be 1) In netdiffuseR is as easy as doing the following: # fakedata In this type of models we usually have the following Now we shuffle toas, so that is random # Resetting TOAs (now will be completely random)ĭiffnet.toa(net) <- sample(diffnet.toa(net), nnodes(net), TRUE) # H0: E - E = 0 (no structure dependency) Statistic = function(x) mean(threshold(x), na.rm = TRUE), To use this test, netdiffuseR has the struct_test function.īasically it simulates networks with the same density and computes a particular statistic every time, generating an EDF (Empirical Distribution Function) under the Null hyphothesis (p-values).Moran(medInnovationsDiffNet$cumadopt, W) # $observed # We get the element-wise <- And then compute moran W <- approx_geodesic(medInnovationsDiffNet$graph]) What happens under the hood is: # For each time point we compute the geodesic distances matrix The summary.diffnet method already runs Moran’s for you. NetdiffuseR = netdiffuseR::approx_geodesic(dgc),
Ig <- igraph::graph_from_adjacency_matrix(dgc) This could be faster (if you only care up to n steps) than igraph or sns: # Extracting the large adjacency matrix (stacked)ĭgc <- diag_expand(medInnovationsDiffNet$graph) geodist is a useful command in Stata that helps you to find the distance between two cities/locations, the nearest location from your target city/location, and the number of cities/location within a certain radius. NetdiffuseR has a function to do so, the approx_geodesic function which, using graph powers, computes the shortest path up to n steps. In the case of sparse matrices, and furthermore, in the presence of structural holes it is more convenient to calculate the distance matrix taking this into account. One approach is to use the geodesic (shortes path length) matrix to account for indirect influence.
0 kommentar(er)
