regression | R Documentation |
This function provides an implementation of the multiple regression quadratic assignment procedure (MRQAP) for both one-mode and two-mode network linear models. It offers several advantages:
it works with combined graph/network objects such as igraph and network objects by constructing the various dependent and independent matrices for the user.
it uses a more intuitive formula-based system for specifying the model, with several ways to specify how nodal attributes should be handled.
it can handle categorical variables (factors/characters) and interactions intuitively, naming the reference variable where appropriate.
it relies on {furrr}
for parallelising
and {progressr}
for reporting progress to the user,
which can be useful when many simulations are required.
results are {broom}
-compatible,
with tidy()
and glance()
reports to facilitate comparison
with results from different models.
Note that a t- or z-value is always used as the test statistic,
and properties of the dependent network
– modes, directedness, loops, etc –
will always be respected in permutations and analysis.
net_regression(
formula,
.data,
method = c("qap", "qapy"),
times = 1000,
strategy = "sequential",
verbose = FALSE
)
formula |
A formula describing the relationship being tested. Several additional terms are available to assist users investigate the effects they are interested in. These include:
|
.data |
A manynet-consistent network.
See e.g. |
method |
A method for establishing the null hypothesis. Note that "qap" uses Dekker et al's (2007) double semi-partialling technique, whereas "qapy" permutes only the $y$ variable. "qap" is the default. |
times |
Integer indicating number of simulations used for quantile estimation.
(Relevant to the null hypothesis test only -
the analysis itself is unaffected by this parameter.)
Note that, as for all Monte Carlo procedures, convergence is slower for more
extreme quantiles.
By default, |
strategy |
If |
verbose |
Whether the function should report on its progress.
By default FALSE.
See |
Krackhardt, David. 1988. “Predicting with Networks: Nonparametric Multiple Regression Analysis of Dyadic Data.” Social Networks 10(4):359–81. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1016/0378-8733(88)90004-4")}.
Dekker, David, David Krackhard, and Tom A. B. Snijders. 2007. “Sensitivity of MRQAP tests to collinearity and autocorrelation conditions.” Psychometrika 72(4): 563-581. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1007/s11336-007-9016-1")}.
Other models:
test_distributions
,
tests
networkers <- ison_networkers %>% to_subgraph(Discipline == "Sociology")
model1 <- net_regression(weight ~ ego(Citations) + alter(Citations) + sim(Citations),
networkers, times = 20)
# Should be run many more `times` for publication-ready results
tidy(model1)
glance(model1)
plot(model1)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.