traning_zone/modele_zone/modeles_gridsearch.py

from sklearn.model_selection import GridSearchCV
from sklearn.metrics import accuracy_score, balanced_accuracy_score, classification_report
from sklearn.model_selection import KFold

import warnings
warnings.filterwarnings("ignore")

import joblib
import os
import json



# dictionnaire des hyperparamètres
parameters =  { "GradientBoostingClassifier": {
        "learning_rate": [0.01, 0.1, 0.5],
        "n_estimators": [50, 100, 200],
        "max_depth": [3, 5, 7]
    },
    "XGBClassifier": {
        "learning_rate": [0.01, 0.1, 0.5],
        "n_estimators": [50, 100, 200],
        "max_depth": [3, 5, 7]
    },
    "naive_bayes": {
        "alpha": [0.1, 0.5, 1.0]
    },
    "LogisticRegression": {
        "penalty": ['l1', 'l2'],
        "C": [0.1, 1, 10]
    },
    "RandomForestClassifier": {
        "n_estimators": [50, 100, 200],
        "max_depth": [None, 5, 10],
        "min_samples_split": [2, 5, 10]
    },
    "svm": {
        "C": [0.1, 1, 10],
        "kernel": ['linear', 'rbf'],
        "gamma": ['scale', 'auto']
    }
}



# fonction de validation croisée
def traning_gridsearch(X_train, Y_train, X_test, Y_test, Model, name,classe, cv = 5):

    try :
        os.mkdir(f'traning_zone/mini_modèles_w/{classe}')
        try : 
            os.mkdir(f'traning_zone/mini_modèles_w/{classe}/{name}')
        except FileExistsError:
             pass 
    except FileExistsError:
        try : 
            os.mkdir(f'traning_zone/mini_modèles_w/{classe}/{name}')
        except FileExistsError:
            pass

    kfold = KFold(n_splits=cv, shuffle=True, random_state=99999)


    grid_search = GridSearchCV(Model, parameters[name], scoring="accuracy",cv=kfold)
        
    grid_search.fit(X_train, Y_train)
        
    best_params =grid_search.best_params_
    best_score = grid_search.best_score_
    best_model = grid_search.best_estimator_

    with open(f"traning_zone/mini_modèles_w/{classe}/{name}/best_score.json", "w") as jsonfile:
        json.dump(best_score, jsonfile)

    with open(f"traning_zone/mini_modèles_w/{classe}/{name}/best_params.json", "w") as jsonfile:
        json.dump(best_params, jsonfile)

    Y_pred = best_model.predict(X_test)

    score = balanced_accuracy_score(Y_test, Y_pred)

    with open(f"traning_zone/mini_modèles_w/{classe}/{name}/balanced_accuracy_score_score.json", "w") as jsonfile:
        json.dump(score, jsonfile)
    
    scores = classification_report(Y_test, Y_pred)

    with open(f"traning_zone/mini_modèles_w/{classe}/{name}/classification_report_score.json", "w") as jsonfil:
        json.dump(scores, jsonfil)
    
    return