CreateSimpleTrainingPlan Class Reference
Last update: 16.07.2025This class is used to create training plans for the ScalarMOP competition. It sets up data according to an sample analysis to allow efficient training of the samples. It provides functions to access and cleanup the prepared data. More...
Public Member Functions | |
| cleanup (data_handler::DataHandlerBase datahandler, std::vector< uint64_t > output_map) | |
| Clean-up the datahandler once the data is not needed anymore. More... | |
| SimpleTrainingPlan | compute (data_handler::DataHandlerBase datahandler, std::vector< uint64_t > output_map, bool use_incompletes=true, bool subspace_filtering=true, bool input_correlation_filter=true, number maximum_input_correlation=0.9) |
| Generates a SimpleTrainingPlan from the classes settings Instances of this class have the ability to cache adapted data inside the datahandler such that the resulting training plan can be used in an efficient way. This is triggered by the outputs for the training. More... | |
| CreateSimpleTrainingPlan () | |
| Constructor. | |
| Matrix | getInputs (data_handler::DataHandlerBase datahandler, std::vector< uint64_t > output_map) |
| Return the input matrix used by the training plan CreateSimpleTrainingPlan::compute may cache efficient input matrices for the given outputs. This function provides access to the input matrices used for the given outputs. More... | |
| initialize (data_handler::DataHandlerBase datahandler) | |
| Initializes and precalculates the cached plan. | |
| IndexVector | unusedSamples (std::vector< uint64_t > output_map, bool use_incompletes) |
| Returns the indices of samples that were left out of the training process for the given output_map. More... | |
Public Attributes | |
| ParameterImportanceVector | input_importances |
| The importances for the mops input variables. | |
| string | mop_ident |
| int | number_of_folds |
| The number of folds for a k-fold training. | |
| TrainingPlanType | training |
| The type of training plan to generate. | |
Detailed Description
This class is used to create training plans for the ScalarMOP competition. It sets up data according to an sample analysis to allow efficient training of the samples. It provides functions to access and cleanup the prepared data.
Member Function Documentation
◆ cleanup()
| cleanup | ( | data_handler::DataHandlerBase | datahandler, |
| std::vector< uint64_t > | output_map | ||
| ) |
Clean-up the datahandler once the data is not needed anymore.
- Parameters
-
datahandler The datahandler that keeps the data output_map The outputs the datahandler keeps additional data for
◆ compute()
| SimpleTrainingPlan compute | ( | data_handler::DataHandlerBase | datahandler, |
| std::vector< uint64_t > | output_map, | ||
| bool | use_incompletes = true, |
||
| bool | subspace_filtering = true, |
||
| bool | input_correlation_filter = true, |
||
| number | maximum_input_correlation = 0.9 |
||
| ) |
Generates a SimpleTrainingPlan from the classes settings Instances of this class have the ability to cache adapted data inside the datahandler such that the resulting training plan can be used in an efficient way. This is triggered by the outputs for the training.
- Parameters
-
datahandler The Datahandler that stores the input/output matrices. output_map The output to create the training plan for use_incompletes The setting if incomplete designs should be used subspace_filtering Optional subspace filtering input_correlation_filter Optional input correlation filtering maximum_input_correlation Maximum input correlation for optional filtering
- Returns
- A training plan
◆ getInputs()
| Matrix getInputs | ( | data_handler::DataHandlerBase | datahandler, |
| std::vector< uint64_t > | output_map | ||
| ) |
Return the input matrix used by the training plan CreateSimpleTrainingPlan::compute may cache efficient input matrices for the given outputs. This function provides access to the input matrices used for the given outputs.
- Parameters
-
datahandler The datahandler where the data is located output_map The output_map used
- Returns
- A reference to the input matrix for the given output_map
◆ unusedSamples()
| IndexVector unusedSamples | ( | std::vector< uint64_t > | output_map, |
| bool | use_incompletes | ||
| ) |
Returns the indices of samples that were left out of the training process for the given output_map.
- Parameters
-
output_map The outputs in question for the training use_incompletes The setting if incomplete designs should be used
- Returns
- An ordered vector of indices of samples that were omitted during training.
