This vignette illustrates the standard use of the
PLNnetwork
function and the methods accompanying the R6
Classes PLNnetworkfamily
and
PLNnetworkfit
.
The packages required for the analysis are PLNmodels plus some others for data manipulation and representation:
We illustrate our point with the trichoptera data set, a full description of which can be found in the corresponding vignette. Data preparation is also detailed in the specific vignette.
The trichoptera
data frame stores a matrix of counts
(trichoptera$Abundance
), a matrix of offsets
(trichoptera$Offset
) and some vectors of covariates
(trichoptera$Wind
, trichoptera$Temperature
,
etc.)
The network model for multivariate count data that we introduce in Chiquet, Robin, and Mariadassou (2019) is a variant of the Poisson Lognormal model of Aitchison and Ho (1989), see the PLN vignette as a reminder. Compare to the standard PLN model we add a sparsity constraint on the inverse covariance matrix \({\boldsymbol\Sigma}^{-1}\triangleq \boldsymbol\Omega\) by means of the \(\ell_1\)-norm, such that \(\|\boldsymbol\Omega\|_1 < c\). PLN-network is the equivalent of the sparse multivariate Gaussian model (Banerjee, Ghaoui, and d’Aspremont 2008) in the PLN framework. It relates some \(p\)-dimensional observation vectors \(\mathbf{Y}_i\) to some \(p\)-dimensional vectors of Gaussian latent variables \(\mathbf{Z}_i\) as follows \[\begin{equation} \begin{array}{rcl} \text{latent space } & \mathbf{Z}_i \sim \mathcal{N}\left({\boldsymbol\mu},\boldsymbol\Omega^{-1}\right) & \|\boldsymbol\Omega\|_1 < c \\ \text{observation space } & Y_{ij} | Z_{ij} \quad \text{indep.} & Y_{ij} | Z_{ij} \sim \mathcal{P}\left(\exp\{Z_{ij}\}\right) \end{array} \end{equation}\]
The parameter \({\boldsymbol\mu}\) corresponds to the main effects and the latent covariance matrix \(\boldsymbol\Sigma\) describes the underlying structure of dependence between the \(p\) variables.
The \(\ell_1\)-penalty on \(\boldsymbol\Omega\) induces sparsity and selection of important direct relationships between entities. Hence, the support of \(\boldsymbol\Omega\) correspond to a network of underlying interactions. The sparsity level (\(c\) in the above mathematical model), which corresponds to the number of edges in the network, is controlled by a penalty parameter in the optimization process sometimes referred to as \(\lambda\). All mathematical details can be found in Chiquet, Robin, and Mariadassou (2019).
Just like PLN, PLN-network generalizes to a formulation close to a multivariate generalized linear model where the main effect is due to a linear combination of \(d\) covariates \(\mathbf{x}_i\) and to a vector \(\mathbf{o}_i\) of \(p\) offsets in sample \(i\). The latent layer then reads \[\begin{equation} \mathbf{Z}_i \sim \mathcal{N}\left({\mathbf{o}_i + \mathbf{x}_i^\top\mathbf{B}},\boldsymbol\Omega^{-1}\right), \qquad \|\boldsymbol\Omega\|_1 < c , \end{equation}\] where \(\mathbf{B}\) is a \(d\times p\) matrix of regression parameters.
Regularization via sparsification of \(\boldsymbol\Omega\) and visualization of the consecutive network is the main objective in PLN-network. To reach this goal, we need to first estimate the model parameters. Inference in PLN-network focuses on the regression parameters \(\mathbf{B}\) and the inverse covariance \(\boldsymbol\Omega\). Technically speaking, we adopt a variational strategy to approximate the \(\ell_1\)-penalized log-likelihood function and optimize the consecutive sparse variational surrogate with an optimization scheme that alternates between two step
More technical details can be found in Chiquet, Robin, and Mariadassou (2019)
In the package, the sparse PLN-network model is adjusted with the
function PLNnetwork
, which we review in this section. This
function adjusts the model for a series of value of the penalty
parameter controlling the number of edges in the network. It then
provides a collection of objects with class PLNnetworkfit
,
corresponding to networks with different levels of density, all stored
in an object with class PLNnetworkfamily
.
PLNnetwork
finds an hopefully appropriate set of
penalties on its own. This set can be controlled by the user, but use it
with care and check details in ?PLNnetwork
. The collection
of models is fitted as follows:
##
## Initialization...
## Adjusting 30 PLN with sparse inverse covariance estimation
## Joint optimization alternating gradient descent and graphical-lasso
## sparsifying penalty = 3.603979
sparsifying penalty = 3.328891
sparsifying penalty = 3.074799
sparsifying penalty = 2.840102
sparsifying penalty = 2.62332
sparsifying penalty = 2.423084
sparsifying penalty = 2.238132
sparsifying penalty = 2.067297
sparsifying penalty = 1.909502
sparsifying penalty = 1.763752
sparsifying penalty = 1.629126
sparsifying penalty = 1.504776
sparsifying penalty = 1.389918
sparsifying penalty = 1.283827
sparsifying penalty = 1.185833
sparsifying penalty = 1.09532
sparsifying penalty = 1.011715
sparsifying penalty = 0.9344916
sparsifying penalty = 0.8631626
sparsifying penalty = 0.7972782
sparsifying penalty = 0.7364226
sparsifying penalty = 0.6802121
sparsifying penalty = 0.6282921
sparsifying penalty = 0.5803351
sparsifying penalty = 0.5360386
sparsifying penalty = 0.4951232
sparsifying penalty = 0.4573309
sparsifying penalty = 0.4224232
sparsifying penalty = 0.39018
sparsifying penalty = 0.3603979
## Post-treatments
## DONE!
Note the use of the formula
object to specify the model,
similar to the one used in the function PLN
.
PLNnetworkfamily
The network_models
variable is an R6
object
with class PLNnetworkfamily
, which comes with a couple of
methods. The most basic is the show/print
method, which
sends a very basic summary of the estimation process:
## --------------------------------------------------------
## COLLECTION OF 30 POISSON LOGNORMAL MODELS
## --------------------------------------------------------
## Task: Network Inference
## ========================================================
## - 30 penalties considered: from 0.3603979 to 3.603979
## - Best model (greater BIC): lambda = 0.495
## - Best model (greater EBIC): lambda = 0.736
One can also easily access the successive values of the criteria in the collection
param | nb_param | loglik | BIC | ICL | n_edges | EBIC | pen_loglik | density | stability |
---|---|---|---|---|---|---|---|---|---|
3.603979 | 34 | -1287.396 | -1353.557 | -2799.494 | 0 | -1353.557 | -1293.854 | 0 | NA |
3.328891 | 34 | -1278.861 | -1345.022 | -2780.471 | 0 | -1345.022 | -1285.185 | 0 | NA |
3.074799 | 34 | -1270.373 | -1336.534 | -2760.494 | 0 | -1336.534 | -1276.575 | 0 | NA |
2.840102 | 34 | -1264.050 | -1330.211 | -2753.319 | 0 | -1330.211 | -1270.022 | 0 | NA |
2.623320 | 34 | -1254.302 | -1320.463 | -2721.767 | 0 | -1320.463 | -1260.253 | 0 | NA |
2.423084 | 34 | -1244.592 | -1310.753 | -2686.611 | 0 | -1310.753 | -1250.555 | 0 | NA |
A diagnostic of the optimization process is available via the
convergence
field:
param | nb_param | status | backend | iterations | objective | convergence | |
---|---|---|---|---|---|---|---|
out | 3.603979 | 34 | 4 | nlopt | 7.000000 | 1287.396062 | 0.000005 |
elt | 3.328891 | 34 | 3 | nlopt | 7.000000 | 1278.860878 | 0.000004 |
elt.1 | 3.074799 | 34 | 4 | nlopt | 7.000000 | 1270.372942 | 0.000002 |
elt.2 | 2.840102 | 34 | 4 | nlopt | 8.000000 | 1264.050498 | 0.000006 |
elt.3 | 2.623320 | 34 | 4 | nlopt | 7.000000 | 1254.302454 | 0.000001 |
elt.4 | 2.423084 | 34 | 3 | nlopt | 7.000000 | 1244.592143 | 0.000000 |
An nicer view of this output comes with the option “diagnostic” in
the plot
method:
By default, the plot
method of
PLNnetworkfamily
displays evolution of the criteria
mentioned above, and is a good starting point for model selection:
Note that we use the original definition of the BIC/ICL criterion
(\(\texttt{loglik} -
\frac{1}{2}\texttt{pen}\)), which is on the same scale as the
log-likelihood. A popular
alternative consists in using \(-2\texttt{loglik} + \texttt{pen}\) instead.
You can do so by specifying reverse = TRUE
:
In this case, the variational lower bound of the log-likelihood is
hopefully strictly increasing (or rather decreasing if using
reverse = TRUE
) with a lower level of penalty (meaning more
edges in the network). The same holds true for the penalized counterpart
of the variational surrogate. Generally, smoothness of these criteria is
a good sanity check of optimization process. BIC and its
extended-version high-dimensional version EBIC are classically used for
selecting the correct amount of penalization with sparse estimator like
the one used by PLN-network. However, we will consider later a more
robust albeit more computationally intensive strategy to chose the
appropriate number of edges in the network.
To pursue the analysis, we can represent the coefficient path (i.e.,
value of the edges in the network according to the penalty level) to see
if some edges clearly come off. An alternative and more intuitive view
consists in plotting the values of the partial correlations along the
path, which can be obtained with the options corr = TRUE
.
To this end, we provide the S3 function
coefficient_path
:
To select a network with a specific level of penalty, one uses the
getModel(lambda)
S3 method. We can also extract the best
model according to the BIC or EBIC with the method
getBestModel()
.
model_pen <- getModel(network_models, network_models$penalties[20]) # give some sparsity
model_BIC <- getBestModel(network_models, "BIC") # if no criteria is specified, the best BIC is used
An alternative strategy is to use StARS (Liu, Roeder, and Wasserman 2010), which performs resampling to evaluate the robustness of the network along the path of solutions in a similar fashion as the stability selection approach of Meinshausen and Bühlmann (2010), but in a network inference context.
Resampling can be computationally demanding but is easily
parallelized: the function stability_selection
integrates
some features of the future package to perform parallel
computing. We set our plan to speed the process by relying on 2
workers:
We first invoke stability_selection
explicitly for
pedagogical purpose. In this case, we need to build our sub-samples
manually:
n <- nrow(trichoptera)
subs <- replicate(10, sample.int(n, size = n/2), simplify = FALSE)
stability_selection(network_models, subsamples = subs)
##
## Stability Selection for PLNnetwork:
## subsampling: ++++++++++
Requesting ‘StARS’ in gestBestmodel
automatically
invokes stability_selection
with 20 sub-samples, if it has
not yet been run.
When “StARS” is requested for the first time,
getBestModel
automatically calls the method
stability_selection
with the default parameters. After the
first call, the stability path is available from the plot
function:
When you are done, do not forget to get back to the standard sequential plan with future.
PLNnetworkfit
The variables model_BIC
, model_StARS
and
model_pen
are other R6Class
objects with class
PLNnetworkfit
. They all inherits from the class
PLNfit
and thus own all its methods, with a couple of
specific one, mostly for network visualization purposes. Most fields and
methods are recalled when such an object is printed:
## Poisson Lognormal with sparse inverse covariance (penalty = 1.39)
## ==================================================================
## nb_param loglik BIC ICL n_edges EBIC pen_loglik density
## 35 -1200.764 -1268.87 -2584.115 1 -1271.327 -1205.678 0.007
## ==================================================================
## * Useful fields
## $model_par, $latent, $latent_pos, $var_par, $optim_par
## $loglik, $BIC, $ICL, $loglik_vec, $nb_param, $criteria
## * Useful S3 methods
## print(), coef(), sigma(), vcov(), fitted()
## predict(), predict_cond(), standard_error()
## * Additional fields for sparse network
## $EBIC, $density, $penalty
## * Additional S3 methods for network
## plot.PLNnetworkfit()
The plot
method provides a quick representation of the
inferred network, with various options (either as a matrix, a graph, and
always send back the plotted object invisibly if users needs to perform
additional analyses).
## IGRAPH 3df1d29 UNW- 17 1 --
## + attr: name (v/c), label (v/c), label.cex (v/n), size (v/n),
## | label.color (v/c), frame.color (v/l), weight (e/n), color (e/c),
## | width (e/n)
## + edge from 3df1d29 (vertex names):
## [1] Hfo--Hsp
We can finally check that the fitted value of the counts – even with sparse regularization of the covariance matrix – are close to the observed ones:
data.frame(
fitted = as.vector(fitted(model_StARS)),
observed = as.vector(trichoptera$Abundance)
) %>%
ggplot(aes(x = observed, y = fitted)) +
geom_point(size = .5, alpha =.25 ) +
scale_x_log10(limits = c(1,1000)) +
scale_y_log10(limits = c(1,1000)) +
theme_bw() + annotation_logticks()