调研Pycaret特征工程

DataTypes_Auto_infer

bool类型转category
查int64和float64中的id列
处理这种情况：pandas把int的列转成了float，因为nan
对于int64的的变量，视基数情况转成float64或类别变量
float64，如果基数=2，转OHE

    # if there are inf or -inf then replace them with NaN
    data.replace([np.inf,-np.inf],np.NaN,inplace=True)

    #remove columns with duplicate name 
    data = data.loc[:,~data.columns.duplicated()]

from pyod.models.knn import KNN
from pyod.models.iforest import IForest
from pyod.models.pca import PCA as PCA_od

# Outlier
class Outlier(BaseEstimator,TransformerMixin):
  '''
    - Removes outlier using ABOD,KNN,IFO,PCA & HOBS using hard voting
    - Only takes numerical / One Hot Encoded features
  '''

  def __init__(self,target,contamination=.20, random_state=42, methods=['knn','iso','pca']):
    self.target = target
    self.contamination = contamination
    self.random_state = random_state
    self.methods = methods

    return(None)

  def fit(self,data,y=None):
    #self.abod.fit(data)
    return(None)

  def transform(self,data,y=None):
    return(data)

  def fit_transform(self,dataset,y=None):

    # dummify if there are any obects 
    if len(dataset.select_dtypes(include='object').columns)>0:
      self.dummy = Dummify(self.target)
      data = self.dummy.fit_transform(dataset)
    else:
      data = dataset.copy()

    # reduce data size for faster processing
    # try:
    #   data = data.astype('float16')
    # except:
    #   None

    if 'knn' in self.methods:
      self.knn = KNN