High level function to correct the computed scores in a hierarchy according to the chosen ensemble algorithm.
Do.TPR.DAG(
threshold = seq(from = 0.1, to = 0.9, by = 0.1),
weight = seq(from = 0.1, to = 0.9, by = 0.1),
kk = 5,
folds = 5,
seed = 23,
norm = TRUE,
norm.type = NULL,
positive = "children",
bottomup = "threshold.free",
topdown = "HTD",
W = NULL,
parallel = FALSE,
ncores = 1,
n.round = 3,
f.criterion = "F",
metric = NULL,
recall.levels = seq(from = 0.1, to = 1, by = 0.1),
compute.performance = FALSE,
flat.file = flat.file,
ann.file = ann.file,
dag.file = dag.file,
flat.dir = flat.dir,
ann.dir = ann.dir,
dag.dir = dag.dir,
hierScore.dir = hierScore.dir,
perf.dir = perf.dir
)range of threshold values to be tested in order to find the best threshold (def: from:0.1,
to:0.9, by:0.1).
The denser the range is, the higher the probability to find the best threshold is, but obviously the execution time will be higher.
Set this parameter only for the thresholded variants; for the threshold-free variants,
the parameter threshold is automatically set to zero.
range of weight values to be tested in order to find the best weight (def: from:0.1, to:0.9, by:0.1).
The denser the range is, the higher the probability to find the best threshold is, but obviously the execution time will be higher.
Set this parameter only for the weighted variants; for the weight-free variants,
the parameter weight is automatically set to zero.
number of folds of the cross validation (def: kk=5) on which tuning the parameters threshold, weight and
tau of the parametric variants of the hierarchical ensemble algorithms. For the non-parametric variants
(i.e. if bottomup = threshold.free), the parameter kk is automatically set to zero.
number of folds of the cross validation on which computing the performance metrics averaged across folds (def. 5).
If folds=NULL, the performance metrics are computed one-shot, otherwise the performance metrics are averaged across folds.
If compute.performance is set to FALSE, folds is automatically set to NULL.
initialization seed for the random generator to create folds (def. 23). If NULL folds are generated without seed
initialization. The parameter seed controls both the parameter kk and the parameter folds.
If compute.performance is set to FALSE and bottomup is set to threshold.free, then
seed is automatically set to NULL.
boolean value: should the flat scores matrix be normalized?
TRUE (def.): the flat scores matrix has been already normalized in according to a normalization method;
FALSE: the flat scores matrix has not been normalized yet. See the parameter norm.type to set the on the fly
normalization method to apply among those possible;
can be one of the following three values:
NULL (def.): set norm.type to NULL if and only if the parameter norm is set to TRUE;
MaxNorm: each score is divided for the maximum of each class;
Qnorm: quantile normalization. preprocessCore package is used;
choice of the positive nodes to be considered in the bottom-up strategy. Can be one of the following values:
children (def.): for each node are considered its positive children;
descendants: for each node are considered its positive descendants;
strategy to enhance the flat predictions by propagating the positive predictions from leaves to root. It can be one of the following values:
threshold.free (def.): positive nodes are selected on the basis of the threshold.free strategy (def.);
threshold: positive nodes are selected on the basis of the threshold strategy;
weighted.threshold.free: positive nodes are selected on the basis of the weighted.threshold.free strategy;
weighted.threshold: positive nodes are selected on the basis of the weighted.threshold strategy;
tau: positive nodes are selected on the basis of the tau strategy;
NOTE: tau is only a DESCENS variants. If you use tau strategy you must set the parameter positive=descendants;
vector of weight relative to a single example. If the vector W is not specified (by def. W=NULL), W is a unitary
vector of the same length of the columns' number of the flat scores matrix (root node included). Set W only if topdown=GPAV.
boolean value:
TRUE: execute the parallel implementation of GPAV (GPAV.parallel);
FALSE (def.): execute the sequential implementation of GPAV (GPAV.over.examples);
Use parallel if and only if topdown=GPAV; otherwise set parallel=FALSE.
number of cores to use for parallel execution (def. 8). Set ncores=1 if parallel=FALSE,
otherwise set ncores to the desired number of cores. Use ncores only if topdown=GPAV; otherwise set parallel=1.
number of rounding digits to be applied to the hierarchical scores matrix (def. 3). It is used for choosing
the best threshold on the basis of the best F-measure.
If compute.performance is set to FALSE and bottomup is set to threshold.free, then
n.round is automatically set to NULL.
character. Type of F-measure to be used to select the best F-measure. Two possibilities:
F (def.): corresponds to the harmonic mean between the average precision and recall;
avF: corresponds to the per-example F-score averaged across all the examples;
If compute.performance is set to FALSE and bottomup is set to threshold.free, then
f.criterion is automatically set to NULL.
a string character specifying the performance metric on which to maximize the parametric ensemble variant. It can be one of the following values:
PRC: the parametric ensemble variant is maximized on the basis of AUPRC (AUPRC);
FMAX: the parametric ensemble variant is maximized on the basis of Fmax (Multilabel.F.measure;
NULL: on the threshold.free variant none parameter optimization is needed, since the variant is non-parametric.
So, if bottomup=threshold.free set metric=NULL (def.);
a vector with the desired recall levels (def: from:0.1, to:0.9, by:0.1) to compute the
Precision at fixed Recall level (PXR). If compute.performance=FALSE the parameter recall.levels is automatically set to NULL.
boolean value: should the flat and hierarchical performance (AUPRC, AUROC, PXR,
multilabel F-score) be returned?
FALSE (def.): performance are not computed and just the hierarchical scores matrix is returned;
TRUE: both performance and hierarchical scores matrix are returned;
name of the file containing the flat scores matrix to be normalized or already normalized (without rda extension).
name of the file containing the label matrix of the examples (without rda extension).
name of the file containing the graph that represents the hierarchy of the classes (without rda extension).
relative path where flat scores matrix is stored.
relative path where annotation matrix is stored.
relative path where graph is stored.
relative path where the hierarchical scores matrix must be stored.
relative path where the performance measures must be stored. If compute.performance=FALSE,
the parameter perf.dir is automatically set to NULL.
Two rda files stored in the respective output directories:
Hierarchical Scores Results: a matrix with examples on rows and classes on columns representing the computed hierarchical scores
for each example and for each considered class. It is stored in the hierScore.dir directory;
Performance Measures: flat and hierarchical performance results:
It is stored in the perf.dir directory.
The parametric hierarchical ensemble variants are cross-validated by maximizing in according to the metric
chosen in the parameter metric, that is F-measure (Multilabel.F.measure) or AUPRC (AUPRC).
The function checks if the number of classes between the flat scores matrix and the annotations matrix mismatched. If so, the number of terms of the annotations matrix is shrunk to the number of terms of the flat scores matrix and the corresponding subgraph is computed as well. N.B.: it is supposed that all the nodes of the subgraph are accessible from the root.
We excluded the predictions of the root node in computing all the performances, since it is a dummy node added to the ontology for practical reasons (e.g. some graph-based software may require a single root node to work). However, the root node scores are stored in the hierarchical scores matrix.
# NOT RUN {
data(graph);
data(scores);
data(labels);
tmpdir <- paste0(tempdir(),"/");
save(g, file=paste0(tmpdir,"graph.rda"));
save(L, file=paste0(tmpdir,"labels.rda"));
save(S, file=paste0(tmpdir,"scores.rda"));
dag.dir <- flat.dir <- ann.dir <- tmpdir;
hierScore.dir <- perf.dir <- tmpdir;
dag.file <- "graph";
flat.file <- "scores";
ann.file <- "labels";
threshold <- weight <- 0;
norm.type <- "MaxNorm";
positive <- "children";
bottomup <- "threshold.free";
topdown <- "HTD";
recall.levels <- seq(from=0.25, to=1, by=0.25);
Do.TPR.DAG(threshold=threshold, weight=weight, kk=NULL, folds=NULL, seed=NULL,
norm=FALSE, norm.type=norm.type, positive=positive, bottomup=bottomup, topdown=topdown,
W=NULL, compute.performance=TRUE, parallel=FALSE, ncores=1, n.round=3, f.criterion="F",
metric=NULL, recall.levels=recall.levels, flat.file=flat.file, ann.file=ann.file,
dag.file=dag.file, flat.dir=flat.dir, ann.dir=ann.dir, dag.dir=dag.dir,
hierScore.dir=hierScore.dir, perf.dir=perf.dir);
# }
Run the code above in your browser using DataLab