Transforms a distance matrix into a matrix of weights

Transforms a distance matrix into weights (1/distance.matrix) normalized by the row sums. Used to compute Moran's I values and Moran's Eigenvector Maps. Allows to apply a threshold to the distance matrix before computing the weights.

Usage

weights_from_distance_matrix(
  distance.matrix = NULL,
  distance.threshold = 0
)

Arguments

distance.matrix

Distance matrix. Default: NULL.

distance.threshold

Numeric, positive, in the range of values of distance.matrix. Distances below this value in the distance matrix are set to 0., Default: 0.

Value

A weighted distance matrix.

See also

Other preprocessing: auto_cor(), auto_vif(), case_weights(), default_distance_thresholds(), double_center_distance_matrix(), is_binary(), make_spatial_fold(), make_spatial_folds(), the_feature_engineer()

Examples

data(plants_distance)

y <- weights_from_distance_matrix(
  distance.matrix = plants_distance
)

y[1:5, 1:5]
#>     columns
#> rows            1            2            3            4            5
#>    1 0.000000e+00 8.547100e-05 1.342814e-05 1.625653e-04 5.150753e-03
#>    2 6.139458e-05 0.000000e+00 1.826840e-05 2.632839e-04 2.903626e-05
#>    3 1.663113e-05 3.149888e-05 0.000000e+00 1.875990e-05 1.435260e-05
#>    4 3.743251e-04 8.439839e-04 3.487753e-05 0.000000e+00 1.287678e-04
#>    5 5.128856e-03 4.025125e-05 1.153917e-05 5.568472e-05 0.000000e+00