Different parameters configure different aspects of training/selection/testing. The learning scenarios set many parameters to corresponding default values, and these can again be changed by the user. Therefore the order in which they are specified can be important.
getConfig(model, name)setConfig(model, name, value)
the model
the name
the value
the value of the configuration parameter
displayThis parameter determines the amount of output of you see at the screen: The larger its value is, the more you see. This can help as a progress indication.
scaleIf set to a true value then for every feature in the training data a scaling is calculated so that its values lie in the interval \([0,1]\). The training then is performed using these scaled values and any testing data is scaled transparently as well.
Because SVMs are not scale-invariant any data should be scaled for two main reasons: First that all features have the same weight, and second to assure that the default gamma parameters that liquidSVM provide remain meaningful.
If you do not have scaled the data previously this is an easy option.
threadsThis parameter determines the number of cores used for computing the kernel matrices, the validation error, and the test error.
threads=0 (default) means that all physical cores of your CPU run one thread.
threads=-1 means that all but one physical cores of your CPU run one thread.
partition_choiceThis parameter determines the way the input space is partitioned. This allows larger data sets for which the kernel matrix does not fit into memory.
partition_choice=0 (default) disables partitioning.
partition_choice=6 gives usually highest speed.
partition_choice=5 gives usually the best test error.
grid_choiceThis parameter determines the size of the hyper- parameter grid used during the training phase. Larger values correspond to larger grids. By default, a 10x10 grid is used. Exact descriptions are given in the next section.
adaptivity_controlThis parameter determines, whether an adaptive grid search heuristic is employed. Larger values lead to more aggressive strategies. The default adaptivity_control = 0 disables the heuristic.
random_seedThis parameter determines the seed for the random generator. random_seed = -1 uses the internal timer create the seed. All other values lead to repeatable behavior of the svm.
foldsHow many folds should be used.
Parameters for regression (least-squares, quantile, and expectile)
clippingThis parameter determines whether the decision functions should be clipped at the specified value. The value clipping = -1.0 leads to an adaptive clipping value, whereas clipping = 0 disables clipping.
Parameter for multiclass classification determine the multiclass strategy: mc-type=0 : AvA with hinge loss. mc-type=1 : OvA with least squares loss. mc-type=2 : OvA with hinge loss. mc-type=3 : AvA with least squares loss.
Parameters for Neyman-Pearson Learning
classThe class, the constraint is enforced on.
constraintThe constraint on the false alarm rate. The script actually considers a couple of values around the value of constraint to give the user an informed choice.
For Support Vector Machines two hyperparameters need to be determined:
gamma the bandwith of the kernel
lambda has to be chosen such that neither over- nor underfitting happen. lambda values are the classical regularization parameter in front of the norm term.
liquidSVM has a built-in a cross-validation scheme to calculate validation errors for many values of these hyperparameters and then to choose the best pair. Since there are two parameters this means we consider a two-dimensional grid.
For both parameters either specific values can be given or a geometrically spaced grid can be specified.
gamma_steps, min_gamma, max_gammaspecifies in the interval between min_gamma and max_gamma there should be gamma_steps many values
gammase.g. gammas=c(0.1,1,10,100) will do these four gamma values
lambda_steps, min_lambda, max_lambdaspecifies in the interval between min_lambda and max_lambda there should be lambda_steps many values
lambdase.g. lambdas=c(0.1,1,10,100) will do these four lambda values
c_valuesthe classical term in front of the empirical error term, e.g. c_values=c(0.1,1,10,100) will do these four cost values (basically inverse of lambdas)
Note the min and max values are scaled according the the number of samples, the dimensionality of the data sets, the number of folds used, and the estimated diameter of the data set.
Using grid_choice allows for some general choices of these parameters
|
0 | 1 | 2 |
|
10 | 15 | 20 |
|
10 | 15 | 20 |
|
0.2 | 0.1 | 0.05 |
|
5.0 | 10.0 | 20.0 |
|
0.001 | 0.0001 | 0.00001 |
|
0.01 | 0.01 | 0.01 |
Using negative values of grid_choice we create a grid with listed gamma and lambda values:
|
-1 |
|
c(10.0, 5.0, 2.0, 1.0, 0.5, 0.25, 0.1, 0.05)
|
|
c(1.0, 0.1, 0.01, 0.001, 0.0001, 0.00001, 0.000001, 0.0000001)
|
|
-2 |
|
c(10.0, 5.0, 2.0, 1.0, 0.5, 0.25, 0.1, 0.05)
|
|
c(0.01, 0.1, 1, 10, 100, 1000, 10000)
|
An adaptive grid search can be activated. The higher the values of MAX_LAMBDA_INCREASES and MAX_NUMBER_OF_WORSE_GAMMAS are set the more conservative the search strategy is. The values can be freely modified.
|
1 | 2 |
|
4 | 3 |
|
4 | 3 |
A major issue with SVMs is that for larger sample sizes the kernel matrix does not fit into the memory any more. Classically this gives an upper limit for the class of problems that traditional SVMs can handle without significant runtime increase. Furthermore also the time complexity is at least \(O(n^2)\).
liquidSVM implements two major concepts to circumvent these issues. One is random chunks which is known well in the literature. However we prefer the new alternative of splitting the space into spatial cells and use local SVMs on every cell.
If you specify useCells=TRUE then the sample space \(X\) gets partitioned into a number of cells. The training is done first for cell 1 then for cell 2 and so on. Now, to predict the label for a value \(x\in X\) liquidSVM first finds out to which cell this \(x\) belongs and then uses the SVM of that cell to predict a label for it.
If you run into memory issues turn cells on: \code{useCells=TRUE}
This is quite performant, since the complexity in both time and memore are both \(O(\mbox{CELLSIZE} \times n)\) and this holds both for training as well as testing! It also can be shown that the quality of the solution is comparable, at least for moderate dimensions.
The cells can be configured using the partition_choice:
This gives a partition into random chunks of size 2000
VORONOI=c(1, 2000)
This gives a partition into 10 random chunks
VORONOI=c(2, 10)
This gives a Voronoi partition into cells with radius not larger than 1.0. For its creation a subsample containing at most 50.000 samples is used.
VORONOI=c(3, 1.0, 50000)
This gives a Voronoi partition into cells with at most 2000 samples (approximately). For its creation a subsample containing at most 50.000 samples is used. A shrinking heuristic is used to reduce the number of cells.
VORONOI=c(4, 2000, 1, 50000)
This gives a overlapping regions with at most 2000 samples (approximately). For its creation a subsample containing at most 50.000 samples is used. A stopping heuristic is used to stop the creation of regions if 0.5 * 2000 samples have not been assigned to a region, yet.
VORONOI=c(5, 2000, 0.5, 50000, 1)
This splits the working sets into Voronoi like with PARTITION_TYPE=4. Unlike that case, the centers for the Voronoi partition are found by a recursive tree approach, which in many cases may be faster.
VORONOI=c(6, 2000, 1, 50000, 2.0, 20, 4,)
The first parameter values correspond to NO_PARTITION, RANDOM_CHUNK_BY_SIZE, RANDOM_CHUNK_BY_NUMBER, VORONOI_BY_RADIUS, VORONOI_BY_SIZE, OVERLAP_BY_SIZE
qt, ex: Here the number of considered tau-quantiles/expectiles as well as the considered tau-values are defined. You can freely change these values but notice that the list of tau-values is space-separated!
npl, roc: Here, you define, which weighted classification problems will be considered. The choice is usually a bit tricky. Good luck ...
NPL: WEIGHT_STEPS=10 MIN_WEIGHT=0.001 MAX_WEIGHT=0.5 GEO_WEIGHTS=1ROC: WEIGHT_STEPS=9 MAX_WEIGHT=0.9 MIN_WEIGHT=0.1 GEO_WEIGHTS=0
By specifying groupIds when initializing an SVM samples obtain group ids. This by default also sets FOLDS_KIND to GROUPED. If the latter is the case then samples with the same group id will be put into the same fold at cross validation. This is important if e.g. there are several patients with several measurements each.
The following parameters should only employed by experienced users and are self-explanatory for these:
KERNELspecifies the kernel to use, at the moment either GAUSS_RBF or POISSON
RETRAIN_METHODAfter training on grids and folds there are only solutions on folds. In order to construct a global solution one can either retrain on the whole training data (SELECT_ON_ENTIRE_TRAIN_SET) or the (partial) solutions from the training are kept and combined using voting (SELECT_ON_EACH_FOLD default)
store_solutions_internallyIf this is true (default in all applicable cases) then the solutions of the train phase are stored and can be just reused in the select phase. If you slowly run out of memory during the train phase maybe disable this. However then in the select phase the best models have to be trained again.
For completeness here are some values that usually get set by the learning scenario
SVM_TYPEKERNEL_RULE, SVM_LS_2D, SVM_HINGE_2D, SVM_QUANTILE, SVM_EXPECTILE_2D, SVM_TEMPLATE
LOSS_TYPECLASSIFICATION_LOSS, MULTI_CLASS_LOSS, LEAST_SQUARES_LOSS, WEIGHTED_LEAST_SQUARES_LOSS, PINBALL_LOSS, TEMPLATE_LOSS
VOTE_SCENARIOVOTE_CLASSIFICATION, VOTE_REGRESSION, VOTE_NPL
KERNEL_MEMORY_MODELLINE_BY_LINE, BLOCK, CACHE, EMPTY
FOLDS_KINDBLOCKS, ALTERNATING, RANDOM, STRATIFIED, GROUPED, RANDOM_SUBSET
WS_TYPEFULL_SET, MULTI_CLASS_ALL_VS_ALL, MULTI_CLASS_ONE_VS_ALL, BOOT_STRAP