model training is done

sunnysavita10 · Aug 24, 2024 · 90ce3b6 · 90ce3b6
1 parent 344a9b0
commit 90ce3b6
Show file tree

Hide file tree

Showing 22 changed files with 44,488 additions and 6 deletions.
diff --git a/Artifacts/08_24_2024_13_18_42/data_ingestion/feature_store/NetworkData.csv b/Artifacts/08_24_2024_13_18_42/data_ingestion/feature_store/NetworkData.csv
diff --git a/Artifacts/08_24_2024_13_18_42/data_ingestion/ingested/test.csv b/Artifacts/08_24_2024_13_18_42/data_ingestion/ingested/test.csv
diff --git a/Artifacts/08_24_2024_13_18_42/data_ingestion/ingested/train.csv b/Artifacts/08_24_2024_13_18_42/data_ingestion/ingested/train.csv
diff --git a/Artifacts/08_24_2024_13_18_42/data_transformation/transformed/test.npy b/Artifacts/08_24_2024_13_18_42/data_transformation/transformed/test.npy
diff --git a/Artifacts/08_24_2024_13_18_42/data_transformation/transformed/train.npy b/Artifacts/08_24_2024_13_18_42/data_transformation/transformed/train.npy
diff --git a/Artifacts/08_24_2024_13_18_42/data_transformation/transformed_object/preprocessing.pkl b/Artifacts/08_24_2024_13_18_42/data_transformation/transformed_object/preprocessing.pkl
diff --git a/Artifacts/08_24_2024_13_18_42/data_validation/drift_report/report.yaml b/Artifacts/08_24_2024_13_18_42/data_validation/drift_report/report.yaml
@@ -0,0 +1,93 @@
+Abnormal_URL:
+  drift_status: false
+  p_value: 0.9998299425911609
+DNSRecord:
+  drift_status: false
+  p_value: 1.0
+Domain_registeration_length:
+  drift_status: false
+  p_value: 0.9999998996954222
+Favicon:
+  drift_status: false
+  p_value: 0.9371431479914312
+Google_Index:
+  drift_status: false
+  p_value: 0.9999996521355587
+HTTPS_token:
+  drift_status: false
+  p_value: 0.9970105474352128
+Iframe:
+  drift_status: false
+  p_value: 0.9943908690886974
+Links_in_tags:
+  drift_status: false
+  p_value: 0.9659080032037665
+Links_pointing_to_page:
+  drift_status: false
+  p_value: 1.0
+Page_Rank:
+  drift_status: false
+  p_value: 0.9998299425911609
+Prefix_Suffix:
+  drift_status: false
+  p_value: 0.9999999999598222
+Redirect:
+  drift_status: false
+  p_value: 1.0
+Request_URL:
+  drift_status: false
+  p_value: 0.9925611176019148
+Result:
+  drift_status: false
+  p_value: 1.0
+RightClick:
+  drift_status: false
+  p_value: 1.0
+SFH:
+  drift_status: false
+  p_value: 1.0
+SSLfinal_State:
+  drift_status: false
+  p_value: 0.971456534142284
+Shortining_Service:
+  drift_status: false
+  p_value: 0.9999953843756324
+Statistical_report:
+  drift_status: false
+  p_value: 1.0
+Submitting_to_email:
+  drift_status: false
+  p_value: 0.9844261985874094
+URL_Length:
+  drift_status: false
+  p_value: 0.8979221700912247
+URL_of_Anchor:
+  drift_status: false
+  p_value: 0.9943908690886974
+age_of_domain:
+  drift_status: false
+  p_value: 0.9597114153770513
+double_slash_redirecting:
+  drift_status: false
+  p_value: 0.9739932435377546
+having_At_Symbol:
+  drift_status: false
+  p_value: 0.9860900857001905
+having_IP_Address:
+  drift_status: false
+  p_value: 0.5890956201196135
+having_Sub_Domain:
+  drift_status: false
+  p_value: 0.971456534142284
+on_mouseover:
+  drift_status: false
+  p_value: 0.8979221700912247
+popUpWidnow:
+  drift_status: false
+  p_value: 0.7786069254909116
+port:
+  drift_status: false
+  p_value: 0.98068847421368
+web_traffic:
+  drift_status: false
+  p_value: 0.7247242015659158
diff --git a/Artifacts/08_24_2024_13_22_11/data_ingestion/feature_store/NetworkData.csv b/Artifacts/08_24_2024_13_22_11/data_ingestion/feature_store/NetworkData.csv
diff --git a/Artifacts/08_24_2024_13_22_11/data_ingestion/ingested/test.csv b/Artifacts/08_24_2024_13_22_11/data_ingestion/ingested/test.csv
diff --git a/Artifacts/08_24_2024_13_22_11/data_ingestion/ingested/train.csv b/Artifacts/08_24_2024_13_22_11/data_ingestion/ingested/train.csv
diff --git a/Artifacts/08_24_2024_13_22_11/data_transformation/transformed/test.npy b/Artifacts/08_24_2024_13_22_11/data_transformation/transformed/test.npy
diff --git a/Artifacts/08_24_2024_13_22_11/data_transformation/transformed/train.npy b/Artifacts/08_24_2024_13_22_11/data_transformation/transformed/train.npy
diff --git a/Artifacts/08_24_2024_13_22_11/data_transformation/transformed_object/preprocessing.pkl b/Artifacts/08_24_2024_13_22_11/data_transformation/transformed_object/preprocessing.pkl
diff --git a/Artifacts/08_24_2024_13_22_11/data_validation/drift_report/report.yaml b/Artifacts/08_24_2024_13_22_11/data_validation/drift_report/report.yaml
@@ -0,0 +1,93 @@
+Abnormal_URL:
+  drift_status: false
+  p_value: 0.09511048419639119
+DNSRecord:
+  drift_status: false
+  p_value: 0.9328006214323754
+Domain_registeration_length:
+  drift_status: false
+  p_value: 0.998821814236978
+Favicon:
+  drift_status: false
+  p_value: 1.0
+Google_Index:
+  drift_status: false
+  p_value: 0.8079952786138995
+HTTPS_token:
+  drift_status: false
+  p_value: 0.9413205455808915
+Iframe:
+  drift_status: false
+  p_value: 1.0
+Links_in_tags:
+  drift_status: false
+  p_value: 0.8427303427983684
+Links_pointing_to_page:
+  drift_status: false
+  p_value: 1.0
+Page_Rank:
+  drift_status: false
+  p_value: 1.0
+Prefix_Suffix:
+  drift_status: false
+  p_value: 1.0
+Redirect:
+  drift_status: false
+  p_value: 0.9970105474352128
+Request_URL:
+  drift_status: false
+  p_value: 0.9964705532516113
+Result:
+  drift_status: false
+  p_value: 0.37182939454122055
+RightClick:
+  drift_status: false
+  p_value: 1.0
+SFH:
+  drift_status: false
+  p_value: 0.9999760319468036
+SSLfinal_State:
+  drift_status: false
+  p_value: 0.8807137057443575
+Shortining_Service:
+  drift_status: false
+  p_value: 0.45991477918984003
+Statistical_report:
+  drift_status: false
+  p_value: 1.0
+Submitting_to_email:
+  drift_status: false
+  p_value: 1.0
+URL_Length:
+  drift_status: false
+  p_value: 0.9999998099738984
+URL_of_Anchor:
+  drift_status: false
+  p_value: 0.41112810853494
+age_of_domain:
+  drift_status: false
+  p_value: 0.9999999999997877
+double_slash_redirecting:
+  drift_status: false
+  p_value: 0.8079952786138995
+having_At_Symbol:
+  drift_status: false
+  p_value: 0.9925611176019148
+having_IP_Address:
+  drift_status: false
+  p_value: 1.0
+having_Sub_Domain:
+  drift_status: false
+  p_value: 0.9964705532516113
+on_mouseover:
+  drift_status: false
+  p_value: 0.9985597993583677
+popUpWidnow:
+  drift_status: false
+  p_value: 1.0
+port:
+  drift_status: false
+  p_value: 0.9999999489728584
+web_traffic:
+  drift_status: false
+  p_value: 0.4527621454370569
diff --git a/Artifacts/08_24_2024_13_22_11/model_trainer/trained_model/model.pkl b/Artifacts/08_24_2024_13_22_11/model_trainer/trained_model/model.pkl
diff --git a/networksecurity/components/__pycache__/model_trainer.cpython-310.pyc b/networksecurity/components/__pycache__/model_trainer.cpython-310.pyc
diff --git a/networksecurity/components/model_trainer.py b/networksecurity/components/model_trainer.py
@@ -12,10 +12,78 @@
 from networksecurity.utils.ml_utils.model.estimator import NetworkModel
 from networksecurity.utils.main_utils.utils import save_object,load_object
 from networksecurity.utils.main_utils.utils import load_numpy_array_data
-
+from networksecurity.utils.ml_utils.metric.classification_metric import get_classification_score
 
 
 
 class ModelTrainer:
-    def __init__(self):
-        pass
+
+    def __init__(self,model_trainer_config:ModelTrainerConfig,
+        data_transformation_artifact:DataTransformationArtifact):
+        try:
+            self.model_trainer_config=model_trainer_config
+            self.data_transformation_artifact=data_transformation_artifact
+        except Exception as e:
+            raise NetworkSecurityException(e,sys)
+
+    def perform_hyper_parameter_tunig(self):
+        pass
+
+
+    def train_model(self,x_train,y_train):
+        try:
+            xgb_clf = XGBClassifier()
+            xgb_clf.fit(x_train,y_train)
+            return xgb_clf
+        except Exception as e:
+            raise e
+
+    def initiate_model_trainer(self)->ModelTrainerArtifact:
+        try:
+            train_file_path = self.data_transformation_artifact.transformed_train_file_path
+            test_file_path = self.data_transformation_artifact.transformed_test_file_path
+
+            #loading training array and testing array
+            train_arr = load_numpy_array_data(train_file_path)
+            test_arr = load_numpy_array_data(test_file_path)
+
+            x_train, y_train, x_test, y_test = (
+                train_arr[:, :-1],
+                train_arr[:, -1],
+                test_arr[:, :-1],
+                test_arr[:, -1],
+            )
+
+            model = self.train_model(x_train, y_train)
+            y_train_pred = model.predict(x_train)
+            classification_train_metric =  get_classification_score(y_true=y_train, y_pred=y_train_pred)
+
+            if classification_train_metric.f1_score<=self.model_trainer_config.expected_accuracy:
+                raise Exception("Trained model is not good to provide expected accuracy")
+
+            y_test_pred = model.predict(x_test)
+            classification_test_metric = get_classification_score(y_true=y_test, y_pred=y_test_pred)
+
+
+            #Overfitting and Underfitting
+            diff = abs(classification_train_metric.f1_score-classification_test_metric.f1_score)
+
+            if diff>self.model_trainer_config.overfitting_underfitting_threshold:
+                raise Exception("Model is not good try to do more experimentation.")
+
+            preprocessor = load_object(file_path=self.data_transformation_artifact.transformed_object_file_path)
+
+            model_dir_path = os.path.dirname(self.model_trainer_config.trained_model_file_path)
+            os.makedirs(model_dir_path,exist_ok=True)
+            Network_Model = NetworkModel(preprocessor=preprocessor,model=model)
+            save_object(self.model_trainer_config.trained_model_file_path, obj=Network_Model)
+
+            #model trainer artifact
+
+            model_trainer_artifact = ModelTrainerArtifact(trained_model_file_path=self.model_trainer_config.trained_model_file_path, 
+            train_metric_artifact=classification_train_metric,
+            test_metric_artifact=classification_test_metric)
+            logging.info(f"Model trainer artifact: {model_trainer_artifact}")
+            return model_trainer_artifact
+        except Exception as e:
+            raise NetworkSecurityException(e,sys)
diff --git a/networksecurity/constant/training_pipeline/__pycache__/__init__.cpython-310.pyc b/networksecurity/constant/training_pipeline/__pycache__/__init__.cpython-310.pyc
diff --git a/networksecurity/entity/__pycache__/artifact_entity.cpython-310.pyc b/networksecurity/entity/__pycache__/artifact_entity.cpython-310.pyc
diff --git a/networksecurity/entity/__pycache__/config_entity.cpython-310.pyc b/networksecurity/entity/__pycache__/config_entity.cpython-310.pyc
diff --git a/networksecurity/entity/artifact_entity.py b/networksecurity/entity/artifact_entity.py
@@ -22,6 +22,11 @@ class DataTransformationArtifact:
     transformed_test_file_path: str
 
 @dataclass
+class ClassificationMetricArtifact:
+    f1_score: float
+    precision_score: float
+    recall_score: float
+@dataclass
 class ModelTrainerArtifact:
     trained_model_file_path: str
     train_metric_artifact: ClassificationMetricArtifact
@@ -35,6 +40,3 @@ class ModelEvaluationArtifact:
 class ModelPusherArtifact:
     pass
 
-@dataclass
-class ClassificationMetricArtifact:
-    pass
diff --git a/networksecurity/pipeline/__pycache__/training_pipeline.cpython-310.pyc b/networksecurity/pipeline/__pycache__/training_pipeline.cpython-310.pyc