SimpleTrainingPlan Struct Reference
Last update: 16.07.2025TrainingPlan assigns a in a set of sample each sample a status on how the sample is to be treated during the training of a model A model is trained for a set of outputs. More...
Public Member Functions | |
| virtual std::vector< string > | getInputNames (::data_handler::DataHandlerBase datahandler) override |
| Returns the input names from the datahandler. More... | |
| virtual Matrix | getInputs (::data_handler::DataHandlerBase datahandler) override |
| Returns the inputs fitting the training plan from the datahandler. More... | |
| virtual string | getMOPIdent () override |
| Returns the MOP ident the TrainingPlan is created for. More... | |
| virtual std::vector< string > | getOutputNames (::data_handler::DataHandlerBase datahandler) override |
| Returns the output names from the datahandler. More... | |
| virtual Matrix | getOutputs (::data_handler::DataHandlerBase datahandler) override |
| Returns the outputs fitting the training plan from the datahandler. More... | |
| virtual SampleUsage | getSampleUsageInSet (uint64_t i, uint64_t j) override |
| getSampleUsage accessor of the j-th sample in the i-th training set More... | |
| virtual SampleUsage | getSampleUsageInSetRef (uint64_t i, uint64_t j) override |
| getSampleUsageRef non- accessor of the j-th sample in the i-th training set More... | |
| virtual ParameterImportance | inputImportance (uint64_t i) override |
| parameter returns the ParameterImportance for the i-th parameter More... | |
| virtual ParameterImportance | inputImportanceInSubspace (uint64_t i, uint64_t j, uint64_t o=0) override |
| inputImportanceInSubspace returns true if the input j is in the subspace i for the given o-th output in the output-matrix. More... | |
| virtual bool | isValid (::data_handler::DataHandlerBase datahandler) override |
| Checks if the training plan is valid for the give datahandler. More... | |
| virtual uint64_t | numInputs () override |
| numInputs returns the number of inputs used and defined in this training plan More... | |
| virtual uint64_t | numOutputs () override |
| numOutputs returns the number of outputs used in this training plan More... | |
| virtual uint64_t | numSamplesInSet (uint64_t i) override |
| numSamplesInSet returns the number of elements (samples) in the training set More... | |
| virtual uint64_t | numSets () override |
| numSets returns the number of training sets in the training plan More... | |
| virtual uint64_t | numSubspaces (uint64_t i=0) override |
| numSubspaces returns the number of filtered subspaces available in the Training Plan to the i-th output in the output matrix More... | |
| virtual uint64_t | numTestSamplesInSet (uint64_t i) override |
| numTestSamplesInSet returns the number of elements (samples) in the training set More... | |
| virtual uint64_t | numTrainingSamplesInSet (uint64_t i) override |
| numTrainingSamplesInSet returns the number of elements (samples) in the training set More... | |
| SimpleTrainingPlan | operator= (SimpleTrainingPlan)=default |
| Copy-Assignment. | |
| virtual uint64_t | output (uint64_t i) override |
| parameter returns the index of the i-th output in the output matrix More... | |
| virtual | setInputs (::data_handler::DataHandlerBase datahandler, Matrix inputs) |
| Inserts the inputs fitting the training plan from the datahandler. More... | |
| virtual | setOutputs (::data_handler::DataHandlerBase datahandler, Matrix outputs) |
| Inserts the outputs fitting the training plan from the datahandler. More... | |
| SimpleTrainingPlan ()=default | |
| Constructor. | |
| SimpleTrainingPlan (TrainingPlanVector plan, ParameterImportanceVector params, IndexVector output_map, SubspaceMatrixList subspace_matrix_list, uint64_t num_inputs, string mop_ident=handler::defaultMOPIdent) | |
| Initializing Constructor. More... | |
| SimpleTrainingPlan (SimpleTrainingPlan)=default | |
| Copy-Contructor. | |
Detailed Description
TrainingPlan assigns a in a set of sample each sample a status on how the sample is to be treated during the training of a model A model is trained for a set of outputs.
- if the set is a singleton (one element), the trained model is a scalar one
- if the set has more than one element the output dimension should be consistent for all samples. if it is not consistent, we are dealing with incomplete samples. Either the model itself has then to decide how to handle this or this should be handled before the competition is defined i.e. by the (Scalar)MOP The training plan is not concerned with failed/incomplete inputs. It should always handle a consistent input/output pair
Constructor & Destructor Documentation
◆ SimpleTrainingPlan()
| SimpleTrainingPlan | ( | TrainingPlanVector | plan, |
| ParameterImportanceVector | params, | ||
| IndexVector | output_map, | ||
| SubspaceMatrixList | subspace_matrix_list, | ||
| uint64_t | num_inputs, | ||
| string | mop_ident = handler::defaultMOPIdent |
||
| ) |
Initializing Constructor.
- Parameters
-
plan A collection of lists that describe for each sample if its used for training, testing or is ignored Multiple lists are used for instance in the Leave-One-Out or cross-validation scenario params Defines for each input parameter its importance (if it is ignored, mandatory or optional) output_map Defines how the vectors in the output matrix are mapped to the actual output in the training subspace_matrix_list Defines a list of possible input subspaces for each output (after output mapping). This may be used by model for increased approximation quality. num_inputs The number of inputs for the training mop_dent The mops ident
Member Function Documentation
◆ getInputNames()
|
overridevirtual |
Returns the input names from the datahandler.
- Parameters
-
datahandler The datahandler to get the data from
- Returns
- The input data names
Implements TrainingPlanBase.
◆ getInputs()
|
overridevirtual |
Returns the inputs fitting the training plan from the datahandler.
- Parameters
-
datahandler The datahandler to get the data from
- Returns
- The (scalar) input data
Implements TrainingPlanBase.
◆ getMOPIdent()
|
overridevirtual |
Returns the MOP ident the TrainingPlan is created for.
- Returns
- The ident for the MOP
Implements TrainingPlanBase.
◆ getOutputNames()
|
overridevirtual |
Returns the output names from the datahandler.
- Parameters
-
datahandler The datahandler to get the data from
- Returns
- The output data names
Implements TrainingPlanBase.
◆ getOutputs()
|
overridevirtual |
Returns the outputs fitting the training plan from the datahandler.
- Parameters
-
datahandler The datahandler to get the data from
- Returns
- The (scalar) output data
Implements TrainingPlanBase.
◆ getSampleUsageInSet()
|
overridevirtual |
getSampleUsage accessor of the j-th sample in the i-th training set
- Parameters
-
i Index of the training set j Index of the sample in training set i
- Returns
- The sample at the given indices
Implements TrainingPlanBase.
◆ getSampleUsageInSetRef()
|
overridevirtual |
getSampleUsageRef non- accessor of the j-th sample in the i-th training set
- Parameters
-
i Index of the training set j Index of the sample in training set i
- Returns
- The sample at the given indices
Implements TrainingPlanBase.
◆ inputImportance()
|
overridevirtual |
parameter returns the ParameterImportance for the i-th parameter
- Parameters
-
i The index of the parameter
- Returns
- The ParameterUsage of parameter i
Implements TrainingPlanBase.
◆ inputImportanceInSubspace()
|
overridevirtual |
inputImportanceInSubspace returns true if the input j is in the subspace i for the given o-th output in the output-matrix.
- Parameters
-
i Index of the subspace j Index of the input parameter o The o-th output the input is important to
- Returns
- ParameterImportance of input j in subspace i for the given output
Implements TrainingPlanBase.
◆ isValid()
|
overridevirtual |
Checks if the training plan is valid for the give datahandler.
- Returns
- true iff the training plan is valid
Implements TrainingPlanBase.
◆ numInputs()
|
overridevirtual |
numInputs returns the number of inputs used and defined in this training plan
- Returns
- The number of parameters the training plan is defined on
Implements TrainingPlanBase.
◆ numOutputs()
|
overridevirtual |
numOutputs returns the number of outputs used in this training plan
- Returns
- The number of outputs the training plan is defined on
Implements TrainingPlanBase.
◆ numSamplesInSet()
|
overridevirtual |
numSamplesInSet returns the number of elements (samples) in the training set
- Parameters
-
i The index of the training set
- Returns
- The number of elements in set i
Implements TrainingPlanBase.
◆ numSets()
|
overridevirtual |
numSets returns the number of training sets in the training plan
- Returns
- The number of training sets
Implements TrainingPlanBase.
◆ numSubspaces()
|
overridevirtual |
numSubspaces returns the number of filtered subspaces available in the Training Plan to the i-th output in the output matrix
- Parameters
-
i The index of the output
- Returns
- The number of Subspaces
Implements TrainingPlanBase.
◆ numTestSamplesInSet()
|
overridevirtual |
numTestSamplesInSet returns the number of elements (samples) in the training set
- Parameters
-
i The index of the training set
- Returns
- The number of elements in set i
Implements TrainingPlanBase.
◆ numTrainingSamplesInSet()
|
overridevirtual |
numTrainingSamplesInSet returns the number of elements (samples) in the training set
- Parameters
-
i The index of the training set
- Returns
- The number of elements in set i
Implements TrainingPlanBase.
◆ output()
|
overridevirtual |
parameter returns the index of the i-th output in the output matrix
- Parameters
-
i The index of the parameter
- Returns
- The index of output i in the output matrix
Implements TrainingPlanBase.
◆ setInputs()
|
virtual |
Inserts the inputs fitting the training plan from the datahandler.
- Parameters
-
datahandler The datahandler to set the data in inputs The data to set
◆ setOutputs()
|
virtual |
Inserts the outputs fitting the training plan from the datahandler.
- Parameters
-
datahandler The datahandler to set the data in outputs The data to set
