File model_builder_intermediate.hpp

class IntermediateModelBuilder : public IModelBuilder

#include <model_builder_intermediate.hpp>

An implementation of the class IModelBuilder that stores intermediate representations of rules, which can still be modified when globally optimizing a rule-based model once it has been learned, that are ultimately converted into a final model using another IModelBuilder.

Public Types

typedef std::pair<std::unique_ptr<ConditionList>, std::unique_ptr<IEvaluatedPrediction>> IntermediateRule

The type of a rule, which can still be modified.

typedef std::vector<IntermediateRule>::iterator iterator

An iterator that provides access to the intermediate representations of rules and allows to modify them.

Public Functions

IntermediateModelBuilder(std::unique_ptr<IModelBuilder> modelBuilderPtr)

Parameters:

modelBuilderPtr – An unique pointer to an object of type IModelBuilder that should be used to build the final model

iterator begin()

Returns an iterator to the beginning of the intermediate representations of rules.

Returns:

An iterator to the beginning

iterator end()

Returns an iterator to the end of the intermediate representations of rules.

Returns:

An iterator to the end

void removeLastRule()

Removes the intermediate representation of the last rule.

uint32 getNumRules() const

Returns the total number of rules.

Returns:

The total number of rules

uint32 getNumUsedRules() const

Returns the number of used rules.

Returns:

The number of used rules

virtual void setNumUsedRules(uint32 numUsedRules) override

Sets the number of used rules.

Parameters:

numUsedRules – The number of used rules

virtual void setDefaultRule(std::unique_ptr<IEvaluatedPrediction> &predictionPtr) override

Sets the default rule of the model.

Parameters:

predictionPtr – A reference to an unique pointer of type IEvaluatedPrediction that stores the scores that are predicted by the default rule

virtual void addRule(std::unique_ptr<ConditionList> &conditionListPtr, std::unique_ptr<IEvaluatedPrediction> &predictionPtr) override

Adds a new rule to the model.

Parameters:

• conditionListPtr – A reference to an unique pointer of type ConditionList that stores the rule’s conditions

• predictionPtr – A reference to an unique pointer of type IEvaluatedPrediction that stores the scores that are predicted by the rule

virtual std::unique_ptr<IRuleModel> buildModel() override

Builds and returns the model.

Returns:

An unique pointer to an object of type IRuleModel that has been built

Private Members

const std::unique_ptr<IModelBuilder> modelBuilderPtr_

std::unique_ptr<IEvaluatedPrediction> defaultPredictionPtr_

std::vector<IntermediateRule> intermediateRuleList_

uint32 numUsedRules_