【Python】Pycaret のとりあえずのset up

備忘録。。。

Pycaretでのとりあえずのset up

Feature importanceを確認するためにPCAは行わない設定。

#data setup
s = setup(
    data=aug_data,                          # dataset
    
    target="Outcome",                       # traget
    ignore_features = [”please wite features to ignore”],
    
    log_experiment=True,                    # save log True or Flase
    experiment_name="test",                 # name of log 
    
    train_size = 0.8,                       # ratio for train data
    session_id=0,                           # random seed
    
    categorical_features=["please wite features to use as a numeric features"],         # to use as catecorical features
    numeric_features=["please wite features to use as a numeric features"],             # to use as numeric features
    numeric_imputation = "mean",            # use mean value instead Nan
    
        
    normalize=True,                         # normalize
    normalize_method="robust",              # normalize type "zscore", "minmax", "maxabs", "robust"   
    fix_imbalance=True,                     # corrction imbalacen data default SMOTE
    
    #pca=True,                              # PCA
    #pca_method='linear'                    # 'linear' 'kernel' 'incremental'
    #pca_components=                        #（int/float型, default = 0.99）

    
    
    
    
    #use_gpu=True,   
    #silent=False,                          # confirming the resule of setup
    

    
    #remove_outliers=True,                  #remove_outliers values


    #for regression
    
    #numeric_imputation='constant',      
    #categorical_imputation='constant'",
)