ID_BuildTrainDataSets for assembling data for the IntermittentDemandBootStrapper() function.
ID_BuildTrainDataSets(
MetaData,
data,
Case = 2L,
TargetVariableName = NULL,
DateVariableName = NULL,
GroupingVariables = NULL,
FC_Periods,
TimeUnit = "week",
PowerRate = 0.5,
SampleRate = 5L,
TargetWindowSamples = 5L
)This is the metadata returned from the ID_MetadataGenerator() function
This is your transactional data
Indicate which data constructor method to use
Your target variable names
Your date variable names
Your grouping variables
The number of periods to forecast
The time period unit, such as "day", "week", or "month"
The calculated for determining the total samples is number of records to the power of PowerRate. Then that values is multiplied by the SampleRate. This ensures that a more representative sample is generated across the data set.
The value used to sample from each level of the grouping variables
The number of different targets to utilize for a single random start date
Returns the count modeling data and the size modeling data
Other Feature Engineering Helper:
AutoFourierFeatures(),
ID_MetadataGenerator(),
ID_TrainingDataGenerator2(),
ID_TrainingDataGenerator()