Ajustement des hyperparamètres d'Isolation Forest

Détection d’anomalies en Python

Bekhruz (Bex) Tuychiev

Kaggle Master, Data Science Content Creator

Ajuster la contamination

Pas de méthode arrêtée pour l'ajuster
Il faut s'appuyer sur :
- intuition
- enseignements de l'EDA
- connaissance métier
- attentes business

Exemple d'enquête

Rechercher des enquêtes similaires
Estimer la part des plus pauvres et des plus riches
Mieux vaut chercher que choisir à l'aveugle

Données de ventes Big Mart

import pandas as pd

big_mart = pd.read_csv("big_mart_sales.csv")
big_mart.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 7060 entries, 0 to 7059
Data columns (total 5 columns):
 #   Column            Non-Null Count  Dtype  
 0   weight            7060 non-null   float64
 1   fat_content       7060 non-null   object 
 2   type              7060 non-null   object 
 3   max_retail_price  7060 non-null   float64
 4   sales             7060 non-null   float64
dtypes: float64(3), object(2)

Encoder les catégorielles

big_mart = pd.get_dummies(big_mart)

   weight  max_retail_price      sales  fat_content_low_fat    fat_content_regular
0    9.30          249.8092  3735.1380                    1                      0
1    5.92           48.2692   443.4228                    0                      1
2   17.50          141.6180  2097.2700                    1                      0
3   19.20          182.0950   732.3800                    0                      1
4    8.93           53.8614   994.7052                    1                      0

evaluate_outlier_classifier

def evaluate_outlier_classifier(model, data):
    # Get labels
    labels = model.fit_predict(data)

    # Return inliers
    return data[labels == 0]

evaluate_regressor

from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import root_mean_squared_error

evaluate_regressor

def evaluate_regressor(inliers):
    X = inliers.drop("sales", axis=1) 
    y = inliers[['sales']]


    X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=10)

    lr = LinearRegression()
    lr.fit(X_train, y_train)


    preds = lr.predict(X_test)
    rmse = root_mean_squared_error(y_test, preds)

    return round(rmse, 3)

Ajuster la contamination

contaminations = [0.05, 0.1, 0.2, 0.3]
scores = dict()


for c in contaminations:
    # Instantiate IForest with the current c
    iforest = IForest(contamination=c, random_state=10)


    # Get inliers with the current IForest
    inliers = evaluate_outlier_classifier(iforest, big_mart)


    # Calculate and store RMSE into scores
    scores[c] = evaluate_regressor(inliers)

Analyser le résultat

print(scores)

{0.05: 1148.555, 0.1: 1147.48, 0.2: 1082.307, 0.3: 1029.33}

Ajuster plusieurs hyperparamètres

estimators = [100, 200, 300,]
max_samples = [0.6, 0.8, 1]
scores = dict()

Produit cartésien

from itertools import product

list(product(estimators, max_samples))

[(100, 0.6),
 (100, 0.8),
 (100, 1),
 (200, 0.6),
 (200, 0.8),
 (200, 1),
 (300, 0.6),
 (300, 0.8),
 (300, 1)]

Dans la boucle

estimators = [100, 200, 300,]
max_samples = [0.6, 0.8, 1]
scores = dict()


for e, m in product(estimators, max_samples):

    # Instantiate an IForest
    iforest = IForest(n_estimators=e, max_samples=m, contamination=.3)


    # Get the inliers with the current IForest
    inliers = evaluate_outlier_classifier(iforest, big_mart)

    # Calculate and store RMSE into scores
    scores[(e, m)] = evaluate_regressor(inliers)

Analyse du résultat

print(scores)

{(100, 0.6): 959.398,
 (100, 0.8): 986.056,
 (100, 1): 1195.875,
 (200, 0.6): 947.628,
 (200, 0.8): 933.115,
 (200, 1): 1195.875,
 (300, 0.6): 949.412,
 (300, 0.8): 935.962,
 (300, 1): 1195.875}

Exécution parallèle

# Faster compuation with n_jobs=-1
iforest = IForest(n_estimators=1000, n_jobs=-1)

iforest.fit(big_mart)

Passons à la pratique !

Détection d’anomalies en Python