
 道  michi bench  ·  4 models

  classification  ·  600 rows  ·  5-fold cross-validation  ·  target purchased  
  preparation: numeric: impute median · categorical: impute most_frequent +     
  onehot · standardise (scale-sensitive models only) — fitted inside each fold  
  seed 0  ·  3.9s                                                               
                                                                                
  Results  (ranked by balanced_accuracy)                                        
                                                                                
  model      balanced_accuracy      95% interval   vs leader                 fit
  ──────────────────────────────────────────────────────────────────────────────
  linear                 0.892   0.8599 – 0.9241   leader                   0.4s
  rf                     0.878   0.8323 – 0.9235   tied with leader         0.9s
                                                   (p=0.589)                    
  hist-gbm               0.875     0.83 – 0.9192   tied with leader         1.9s
                                                   (p=0.589)                    
  dummy                    0.5         0.5 – 0.5   worse (p<0.0001)         0.0s
                                                                                
  Verdict  linear scores highest, but rf, hist-gbm are statistically            
  indistinguishable from it at this sample size. Choosing between them on these 
  numbers alone is not supported.                                               
                                                                                
  Findings (1)                                                                  
                                                                                
  warn    dataset  2 model(s) are not statistically distinguishable from linear:
                   rf, hist-gbm                                                 
                                                                                
