| Index | Topics |
| bn.var {bnlearn} | R Documentation |
Structure variability of Bayesian networks
Description
Measure the variability of the structure of a Bayesian network.
Usage
# first and second moments' estimation bn.moments(data, R = 200, m = nrow(data), algorithm, algorithm.args = list(), reduce = NULL, debug = FALSE) # descriptive statistics bn.var(x, method) # Monte Carlo test for entropy bn.var.test(x, method, R, B, debug = FALSE)
Arguments
data |
a data frame containing the variables in the model. |
R |
a positive integer, the number of bootstrap replicates (in bn.moments) or the number of
Monte Carlo samples (in bn.var.test). |
m, B |
a positive integer, the size of each bootstrap (in bn.moments) or Monte Carlo (in
bn.var.test) replicate. |
algorithm |
a character string, the learning algorithm to be applied to the bootstrap replicates. Possible values
are gs, iamb, fast.iamb, inter.iamb, mmpc
and hc. See bnlearn-package and the
documentation of each algorithm for details. |
algorithm.args |
a list of extra arguments to be passed to the learning algorithm. |
x |
a covariance matrix or an object of class mvber.moments (the return value of the
bn.moments function). |
method |
a character string, the label of the statistic used in bn.var or bn.var.test.
Possible values are tvar (total variance), gvar (generalized
variance), nvar (Frobenius matrix norm, which is equivalent to Nagao's
test). |
reduce |
a character string, either first or second. If first all the
arcs with first moment equal to zero are dropped; if if second all the arcs with zero variance
are dropped. |
debug |
a boolean value. If TRUE a lot of debugging output is printed; otherwise the function is
completely silent. |
Value
bn.moments returns an object of class mvber.moments.
bn.var returns a vector of two elements, the observed value of the statistic (named
statistic) and its normalized equivalent (named normalized).
bn.var.test returns an object of class htest.
Note
These functions are experimental implementations of techniques still in development; their form (name, parameters, etc.) will likely change without notice in the future.
Author(s)
Marco Scutari
References
Scutari M (2009). "Structure Variability in Bayesian Networks". ArXiv Statistics - Methodology e-prints. http://arxiv.org/abs/0909.1685.
Examples
## Not run: z = bn.moments(learning.test, algorithm = "gs", R = 100) bn.var(z, method = "tvar") # statistic normalized # 1.29060 0.34416 bn.var.test(z, method = "nvar") # # Squared Frobenius Norm # # data: covariance matrix # nvar = 0.5471, B = 5000, R = 100, p-value < 2.2e-16 # alternative hypothesis: true value is greater than 0 ## End(Not run)
| Index | Topics |
