Retrive threshold levels from the exposure matrix

Thresholds are each vertexes exposure at the time of adoption. Substantively it is the proportion of adopters required for each ego to adopt. (see exposure).

threshold(
  obj,
  toa,
  t0 = min(toa, na.rm = TRUE),
  include_censored = FALSE,
  lags = 0L,
  ...
)

Arguments

obj	Either a \(n\times T\) matrix (eposure to the innovation obtained from exposure) or a diffnet object.
toa	Integer vector. Indicating the time of adoption of the innovation.
t0	Integer scalar. See toa_mat.
include_censored	Logical scalar. When TRUE (default), threshold
lags	Integer scalar. Number of lags to consider when computing thresholds. lags=1 defines threshold as exposure at \(T-1\), where T is time of adoption. levels are not reported for observations adopting in the first time period.
...	Further arguments to be passed to exposure.

Value

A vector of size \(n\) indicating the threshold for each node.

Details

By default exposure is not computed for vertices adopting at the first time period, include_censored=FALSE, as estimating threshold for left censored data may yield biased outcomes.

See also

Threshold can be visualized using plot_threshold

Other statistics: bass, classify_adopters(), cumulative_adopt_count(), dgr(), ego_variance(), exposure(), hazard_rate(), infection(), moran(), struct_equiv(), vertex_covariate_dist()

Examples

# Generating a random graph with random Times of Adoption
set.seed(783)
toa <- sample.int(4, 5, TRUE)
graph <- rgraph_er(n=5, t=max(toa) - min(toa) + 1)

# Computing exposure using Structural Equivalnece
adopt <- toa_mat(toa)
se <- struct_equiv(graph)
#> Warning: no non-missing arguments to max; returning -Inf
#> Warning: no non-missing arguments to max; returning -Inf
#> Warning: no non-missing arguments to max; returning -Inf
se <- lapply(se, function(x) methods::as((x$SE)^(-1), "dgCMatrix"))
expo <- exposure(graph, adopt$cumadopt, alt.graph=se)
#> Warning: The -alt.graph- will be treated as 0/1 graph (value=FALSE).

# Retrieving threshold
threshold(expo, toa)
#>   threshold
#> 1       1.0
#> 2       0.4
#> 3       0.8
#> 4       0.2
#> 5       0.8

# We can do the same by creating a diffnet object
diffnet <- as_diffnet(graph, toa)
threshold(diffnet, alt.graph=se)
#> Warning: The -alt.graph- will be treated as 0/1 graph (value=FALSE).
#>   threshold
#> 1       1.0
#> 2       0.4
#> 3       0.8
#> 4       0.2
#> 5       0.8