StructuralModel {sharp} | R Documentation |
Stability selection in Structural Equation Modelling
Description
Performs stability selection for Structural Equation Models. The underlying arrow selection algorithm (e.g. regularised Structural Equation Modelling) is run with different combinations of parameters controlling the sparsity (e.g. penalty parameter) and thresholds in selection proportions. These two hyper-parameters are jointly calibrated by maximisation of the stability score.
Usage
StructuralModel(
xdata,
adjacency,
residual_covariance = NULL,
Lambda = NULL,
pi_list = seq(0.01, 0.99, by = 0.01),
K = 100,
tau = 0.5,
seed = 1,
n_cat = NULL,
implementation = PenalisedLinearSystem,
resampling = "subsampling",
cpss = FALSE,
PFER_method = "MB",
PFER_thr = Inf,
FDP_thr = Inf,
Lambda_cardinal = 100,
optimisation = c("grid_search", "nloptr"),
n_cores = 1,
output_data = FALSE,
verbose = TRUE,
...
)
Arguments
xdata |
matrix with observations as rows and variables as columns.
Column names must be defined and in line with the row and column names of
|
adjacency |
binary adjacency matrix of the Directed Acyclic Graph (transpose of the asymmetric matrix A in Reticular Action Model notation). The row and column names of this matrix must be defined. |
residual_covariance |
binary and symmetric matrix encoding the nonzero entries in the residual covariance matrix (symmetric matrix S in Reticular Action Model notation). By default, this is the identity matrix (no residual covariance). |
Lambda |
matrix of parameters controlling the level of sparsity in the
underlying feature selection algorithm specified in |
pi_list |
vector of thresholds in selection proportions. If
|
K |
number of resampling iterations. |
tau |
subsample size. Only used if |
seed |
value of the seed to initialise the random number generator and
ensure reproducibility of the results (see |
n_cat |
computation options for the stability score. Default is
|
implementation |
function to use for variable selection. |
resampling |
resampling approach. Possible values are:
|
cpss |
logical indicating if complementary pair stability selection
should be done. For this, the algorithm is applied on two non-overlapping
subsets of half of the observations. A feature is considered as selected if
it is selected for both subsamples. With this method, the data is split
|
PFER_method |
method used to compute the upper-bound of the expected
number of False Positives (or Per Family Error Rate, PFER). If
|
PFER_thr |
threshold in PFER for constrained calibration by error
control. If |
FDP_thr |
threshold in the expected proportion of falsely selected
features (or False Discovery Proportion) for constrained calibration by
error control. If |
Lambda_cardinal |
number of values in the grid of parameters controlling
the level of sparsity in the underlying algorithm. Only used if
|
optimisation |
character string indicating the type of optimisation
method. With |
n_cores |
number of cores to use for parallel computing (see argument
|
output_data |
logical indicating if the input datasets |
verbose |
logical indicating if a loading bar and messages should be printed. |
... |
additional parameters passed to the functions provided in
|
Details
In stability selection, a feature selection algorithm is fitted on
K
subsamples (or bootstrap samples) of the data with different
parameters controlling the sparsity (Lambda
). For a given (set of)
sparsity parameter(s), the proportion out of the K
models in which
each feature is selected is calculated. Features with selection proportions
above a threshold pi are considered stably selected. The stability
selection model is controlled by the sparsity parameter(s) for the
underlying algorithm, and the threshold in selection proportion:
V_{\lambda, \pi} = \{ j: p_{\lambda}(j) \ge \pi \}
In Structural Equation Modelling, "feature" refers to an arrow in the corresponding Directed Acyclic Graph.
These parameters can be calibrated by maximisation of a stability score
(see ConsensusScore
if n_cat=NULL
or
StabilityScore
otherwise) calculated under the null
hypothesis of equiprobability of selection.
It is strongly recommended to examine the calibration plot carefully to
check that the grids of parameters Lambda
and pi_list
do not
restrict the calibration to a region that would not include the global
maximum (see CalibrationPlot
). In particular, the grid
Lambda
may need to be extended when the maximum stability is
observed on the left or right edges of the calibration heatmap. In some
instances, multiple peaks of stability score can be observed. Simulation
studies suggest that the peak corresponding to the largest number of
selected features tend to give better selection performances. This is not
necessarily the highest peak (which is automatically retained by the
functions in this package). The user can decide to manually choose another
peak.
To control the expected number of False Positives (Per Family Error Rate)
in the results, a threshold PFER_thr
can be specified. The
optimisation problem is then constrained to sets of parameters that
generate models with an upper-bound in PFER below PFER_thr
(see
Meinshausen and Bühlmann (2010) and Shah and Samworth (2013)).
Possible resampling procedures include defining (i) K
subsamples of
a proportion tau
of the observations, (ii) K
bootstrap
samples with the full sample size (obtained with replacement), and (iii)
K/2
splits of the data in half for complementary pair stability
selection (see arguments resampling
and cpss
). In
complementary pair stability selection, a feature is considered selected at
a given resampling iteration if it is selected in the two complementary
subsamples.
To ensure reproducibility of the results, the starting number of the random
number generator is set to seed
.
For parallelisation, stability selection with different sets of parameters
can be run on n_cores
cores. Using n_cores > 1
creates a
multisession
. Alternatively,
the function can be run manually with different seed
s and all other
parameters equal. The results can then be combined using
Combine
.
Value
An object of class variable_selection
. A list with:
S |
a matrix of the best stability scores for different parameters controlling the level of sparsity in the underlying algorithm. |
Lambda |
a matrix of parameters controlling the level of sparsity in the underlying algorithm. |
Q |
a matrix of the average number of selected features by the underlying algorithm with different parameters controlling the level of sparsity. |
Q_s |
a matrix of the calibrated number of stably selected features with different parameters controlling the level of sparsity. |
P |
a matrix of calibrated thresholds in selection proportions for different parameters controlling the level of sparsity in the underlying algorithm. |
PFER |
a matrix of upper-bounds in PFER of calibrated stability selection models with different parameters controlling the level of sparsity. |
FDP |
a matrix of upper-bounds in FDP of calibrated stability selection models with different parameters controlling the level of sparsity. |
S_2d |
a matrix of stability scores obtained with different combinations of parameters. Columns correspond to different thresholds in selection proportions. |
PFER_2d |
a matrix of upper-bounds in FDP obtained with different combinations of parameters. Columns correspond to different thresholds in selection proportions. |
FDP_2d |
a matrix of upper-bounds in PFER obtained with different combinations of parameters. Columns correspond to different thresholds in selection proportions. |
selprop |
a matrix of selection proportions.
Columns correspond to predictors from |
Beta |
an array
of model coefficients. Columns correspond to predictors from |
method |
a list with
|
params |
a list with values used for arguments
|
For all matrices and arrays returned, the rows
are ordered in the same way and correspond to parameter values stored in
Lambda
.
References
Bodinier B, Filippi S, Nøst TH, Chiquet J, Chadeau-Hyam M (2023). “Automated calibration for stability selection in penalised regression and graphical models.” Journal of the Royal Statistical Society Series C: Applied Statistics, qlad058. ISSN 0035-9254, doi:10.1093/jrsssc/qlad058, https://academic.oup.com/jrsssc/advance-article-pdf/doi/10.1093/jrsssc/qlad058/50878777/qlad058.pdf.
Meinshausen N, Bühlmann P (2010). “Stability selection.” Journal of the Royal Statistical Society: Series B (Statistical Methodology), 72(4), 417-473. doi:10.1111/j.1467-9868.2010.00740.x.
Shah RD, Samworth RJ (2013). “Variable selection with error control: another look at stability selection.” Journal of the Royal Statistical Society: Series B (Statistical Methodology), 75(1), 55-80. doi:10.1111/j.1467-9868.2011.01034.x.
Jacobucci R, Grimm KJ, McArdle JJ (2016). “Regularized structural equation modeling.” Structural equation modeling: a multidisciplinary journal, 23(4), 555–566. doi:10.1080/10705511.2016.1154793.
See Also
SelectionAlgo
,
Resample
, StabilityScore
Other stability functions:
BiSelection()
,
Clustering()
,
GraphicalModel()
,
VariableSelection()
Examples
oldpar <- par(no.readonly = TRUE)
par(mar = rep(7, 4))
# Data simulation
set.seed(1)
pk <- c(3, 2, 3)
simul <- SimulateStructural(
n = 500,
pk = pk,
nu_between = 0.5,
v_between = 1,
v_sign = 1
)
# Stability selection (using glmnet)
dag <- LayeredDAG(layers = pk)
stab <- StructuralModel(
xdata = simul$data,
adjacency = dag
)
CalibrationPlot(stab)
LinearSystemMatrix(vect = Stable(stab), adjacency = dag)
# Stability selection (using OpenMx)
if (requireNamespace("OpenMx", quietly = TRUE)) {
stab <- StructuralModel(
xdata = simul$data,
implementation = PenalisedOpenMx,
Lambda = seq(50, 500, by = 50),
adjacency = dag
)
CalibrationPlot(stab)
OpenMxMatrix(SelectedVariables(stab), adjacency = dag)
}
## Not run:
# Data simulation with latent variables
set.seed(1)
pk <- c(3, 2, 3)
simul <- SimulateStructural(
n = 500,
pk = pk,
nu_between = 0.5,
v_sign = 1,
v_between = 1,
n_manifest = 3,
ev_manifest = 0.95
)
# Stability selection (using OpenMx)
if (requireNamespace("OpenMx", quietly = TRUE)) {
dag <- LayeredDAG(layers = pk, n_manifest = 3)
penalised <- dag
penalised[, seq_len(ncol(simul$data))] <- 0
stab <- StructuralModel(
xdata = simul$data,
implementation = PenalisedOpenMx,
adjacency = dag,
penalised = penalised,
Lambda = seq(10, 100, by = 20),
K = 10 # to increase for real use
)
CalibrationPlot(stab)
ids_latent <- grep("f", colnames(dag))
OpenMxMatrix(SelectedVariables(stab),
adjacency = dag
)[ids_latent, ids_latent]
}
## End(Not run)
par(oldpar)