API Reference - Others - ModelDatasetCreator

Modifies existing dataset so that it can be used by the models.

Constructors

new()

Creates a ModelDatasetCreator object.

ModelDatasetCreator.new(): ModelDatasetCreatorObject

Functions

setDatasetSplitPercentages()

Set the split percentages for training, validation and testing. The current default values are 0.7 for training and 0.3 for testing.

ModelDatasetCreator:setDatasetSplitPercentages(trainDataPercentage: number, validationDataPercentage: number, testDataPercentage: number)

Parameters:

  • trainDataPercentage: The percentage of dataset to be turned to training data. The value must be between 0 and 1.

  • validationDataPercentage: The percentage of dataset to be turned to validation data. The value must be between 0 and 1.

  • testDataPercentage: The percentage of dataset to be turned to testing data. The value must be between 0 and 1.

setDatasetRandomizationProbability() 

ModelDatasetCreator:setDatasetRandomizationProbability(datasetRandomizationProbability: number)

Parameters:

  • datasetRandomizationProbability: The probability to randomize the positions of each datapoints. The higher the value, the more likely it will be randomized. The value must be between 0 and 1.

randomizeDataset()

Randomizes the each data positions in te dataset

ModelDatasetCreator:randomizeDataset(featureMatrix: matrix, labelVectorOrMatrix: matrix): matrix, matrix

Parameters:

  • featureMatrix: The matrix containing all the data.

  • labelVectorOrMatrix: The matrix containing all the label values related to feature matrix. Optional argument.

Returns:

  • randomizedFeatureMatrix: The matrix containing all the data.

  • randomizedLabelVectorOrMatrix: The matrix containing all the label values related to feature matrix. Only returns if labelVectorOrMatrix is added.

splitDataset()

Predict the values for given data.

ModelDatasetCreator:splitDataset(datasetMatrix): matrix, matrix, matrix

Parameters:

  • datasetMatrix: The feature matrix or label matrix to split.

Returns:

  • trainDatasetMatrix: The dataset for training models.

  • validationDatasetMatrix: The dataset for validating models.

  • testDatasetMatrix: The dataset for testing models.