The Spider Objects

 
 
 
 
   
   CV cross validation object      
   
 


    a=cv(algo,hyperParam) Returns a cv object on algorithm algo using with 
                          given Hyperparameters. 
   
    Possible hyperparameters (with defaults):
     folds=5               - number of folds
     repeats=1             - can do n*CV for reduced variance
     balanced=0            - determines if cv shall be balanced  (same number of
                              positives in each fold)
     store_all=1           - determines if models trained in all folds shall be stored   
     output_train_error=0  - determines wheter to output training error on each fold
                              (else cv error, i.e  test left out fold as default)
     train_on_fold=0       - determines to  test on left out fold and train on the
                              rest (set to true for the opposite).
     store_trialbytrial=0  - whether to store the field a.trialbytrial, the rows of
                              which are [i, foldnumber, y_i, f(x_i)] for each index
                              i of a data point on which the algorithm was tested
  
    Model:
     child                 - stored in child algorithm 
  
    Methods:
     train                 - selfexplanatory
     test                  - selfexplanatory