14  Class Diagrams

Note

This section contains class diagrams for the class system used internally by rtemis. It is intended for developers and advanced users who want to understand the structure of the package, not for most userss.

14.1 Resampling: ResamplerParameters & Resampler

classDiagram
    class ResamplerParameters {
        <<abstract>>
        +character type
        +integer n
        +print(pad) void
        +desc() character
        +$() prop
        +[[]]() prop
    }
    
    class KFoldParams {
        +character|NULL stratify_var
        +integer strat_n_bins
        +vector|NULL id_strat
        +integer seed
        +constructor(n, stratify_var, strat_n_bins, id_strat, seed)
    }
    
    class StratSubParams {
        +integer n
        +double train_p
        +character|NULL stratify_var
        +integer strat_n_bins
        +vector|NULL id_strat
        +integer seed
        +constructor(n, train_p, stratify_var, strat_n_bins, id_strat, seed)
    }
    
    class StratBootParams {
        +character|NULL stratify_var
        +double train_p
        +integer strat_n_bins
        +integer target_length
        +vector|NULL id_strat
        +integer seed
        +constructor(n, stratify_var, train_p, strat_n_bins, target_length, id_strat, seed)
    }
    
    class BootstrapParams {
        +vector|NULL id_strat
        +integer seed
        +constructor(n, id_strat, seed)
    }
    
    class LOOCVParams {
        +constructor(n)
    }
    
    class CustomParams {
        +constructor(n)
    }
    
    class Resampler {
        +character type
        +list resamples
        +ResamplerParameters parameters
        +print() void
        +names() character[]
        +$() resample
        +[[]]() resample
        +desc() character
        +.DollarNames() character[]
    }
    
    ResamplerParameters <|-- KFoldParams
    ResamplerParameters <|-- StratSubParams
    ResamplerParameters <|-- StratBootParams
    ResamplerParameters <|-- BootstrapParams
    ResamplerParameters <|-- LOOCVParams
    ResamplerParameters <|-- CustomParams
    
    Resampler *-- ResamplerParameters : contains
    
    note for ResamplerParameters "Base class for all resampler parameter classes\nProvides common properties: type and n"
    note for Resampler "Main resampling class that stores\nresamples and their parameters"

14.1.1 Class Descriptions

ResamplerParameters (Abstract Base Class)

• Superclass for all resampler parameter classes
• Contains common properties: type (character) and n (integer)
  
• Provides methods for printing, description, and property access

Parameter Subclasses:

• KFoldParams: K-fold cross-validation parameters
• StratSubParams: Stratified subsampling parameters with train/test split ratio
  
• StratBootParams: Stratified bootstrap parameters with target length
• BootstrapParams: Standard bootstrap parameters
• LOOCVParams: Leave-one-out cross-validation parameters (minimal properties)
• CustomParams: Custom resampling parameters (minimal properties)

Resampler (Main Class)

• Stores the actual resamples and their parameters
• Composition relationship with ResamplerParameters
• Provides methods for accessing individual resamples and metadata