“Comment trouver des valeurs aberrantes à Python” Réponses codées

Comment trouver des valeurs aberrantes à Python

# outlier detection function
def outlier(col:str, df_:pd.DataFrame=df, remove:bool=False) -> list:
    """This function calculates the upper and lower fence
    of any column and can also remove from the dataset"""
    q1 = df_[col].quantile(0.25)
    q3 = df_[col].quantile(0.75)
    
    iqr = q3-q1
    lower_fence = q1 - iqr*1.5
    upper_fence = q3 + iqr*1.5

    if remove:
        temp = df_[(df_[col] > lower_fence) & (df_[col] < upper_fence)]
        return temp

    return [lower_fence, upper_fence]