Add files from Desktop Soumya folder

main.py

@@ -0,0 +1,21 @@
+import pandas as pd
+
+df = pd.read_csv("kdd_test.csv")
+
+print("Initial Dataset:")
+print(df.head())
+
+col_to_remove = "urgent"
+
+if col_to_remove in df.columns:
+    df.drop(col_to_remove, axis=1, inplace=True)
+    print(f"\nColumn '{col_to_remove}' removed successfully.")
+else:
+    print(f"\nColumn '{col_to_remove}' not found in dataset.")
+
+
+print("\nDataset after removing the column:")
+print(df.head())
+
+df.to_csv("cleaned_kdd_test.csv", index=False)
+print("\nModified dataset saved as 'cleaned_kdd_test.csv' in the same folder.")

mdltraining.py

@@ -0,0 +1,156 @@
+import pandas as pd
+import numpy as np
+
+from sklearn.preprocessing import StandardScaler, LabelEncoder
+from sklearn.metrics import (
+    accuracy_score,
+    confusion_matrix,
+    classification_report,
+    ConfusionMatrixDisplay
+)
+
+from sklearn.linear_model import LogisticRegression, LinearRegression, SGDClassifier
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.naive_bayes import GaussianNB
+
+import matplotlib.pyplot as plt
+import seaborn as sns
+
+# ===============================================
+# 1. LOAD TRAIN & TEST DATA
+# ===============================================
+train_data = pd.read_csv("kdd_train.csv")
+test_data  = pd.read_csv("kdd_test.csv")
+
+print("Train Dataset Size :", train_data.shape)
+print("Test Dataset Size  :", test_data.shape)
+
+# ===============================================
+# 2. CREATE BINARY TARGET LABEL
+# ===============================================
+train_data["is_attack"] = train_data["labels"].apply(lambda x: 0 if x == "normal" else 1)
+test_data["is_attack"]  = test_data["labels"].apply(lambda x: 0 if x == "normal" else 1)
+
+cat_columns = ["protocol_type", "service", "flag"]
+label_enc = LabelEncoder()
+
+for col in cat_columns:
+    train_data[col] = label_enc.fit_transform(train_data[col])
+    test_data[col] = label_enc.transform(test_data[col])
+
+TARGET = "is_attack"
+FEATURE_COLS = [c for c in train_data.columns if c not in ["is_attack", "labels"]]
+
+# ===============================================
+# 3. SPLIT FEATURES AND TARGET
+# ===============================================
+X_train = train_data[FEATURE_COLS]
+y_train = train_data[TARGET]
+
+X_test  = test_data[FEATURE_COLS]
+y_test  = test_data[TARGET]
+
+# ===============================================
+# 4. FEATURE SCALING
+# ===============================================
+scaler = StandardScaler()
+X_train_scaled = scaler.fit_transform(X_train)
+X_test_scaled  = scaler.transform(X_test)
+
+# ===============================================
+# 5. CORRELATION HEATMAP
+# ===============================================
+num_data = train_data.select_dtypes(include=['number'])
+
+plt.figure(figsize=(14, 11))
+sns.heatmap(num_data.corr(), cmap="viridis", linewidths=0.3)
+plt.title("Correlation Heatmap - Training Dataset")
+plt.show()
+
+# ===============================================
+# 6. MODEL DEFINITIONS
+# ===============================================
+models = {
+    "Linear OLS": LinearRegression(),
+    "SGD Classifier": SGDClassifier(max_iter=2000, tol=1e-4, random_state=42),
+    "Logistic Model": LogisticRegression(max_iter=1500),
+    "Naive Bayes": GaussianNB(),
+    "Decision Tree": DecisionTreeClassifier(random_state=42),
+    "Random Forest": RandomForestClassifier(n_estimators=250, n_jobs=-1, random_state=42)
+}
+
+# ===============================================
+# 7. TRAIN + EVALUATE ALL MODELS
+# ===============================================
+results = {}
+
+for model_name, model in models.items():
+    print(f"\n---- Training Model: {model_name} ----")
+
+    model.fit(X_train_scaled, y_train)
+
+    if model_name == "Linear OLS":
+        y_cont = model.predict(X_test_scaled)
+        y_pred = (y_cont >= 0.5).astype(int)
+    else:
+        y_pred = model.predict(X_test_scaled)
+
+    cm = confusion_matrix(y_test, y_pred)
+    acc = accuracy_score(y_test, y_pred)
+    rep = classification_report(y_test, y_pred, output_dict=True)
+
+    print(f"Accuracy Score : {acc:.4f}")
+    print(f"Recall (Attack): {rep['1']['recall']:.4f}")
+    print("Confusion Matrix:\n", cm)
+
+    results[model_name] = {
+        "cm": cm,
+        "acc": acc,
+        "report": rep
+    }
+
+# ===============================================
+# 8. PLOT CONFUSION MATRICES
+# ===============================================
+rows, cols = 3, 2
+fig, axes = plt.subplots(rows, cols, figsize=(15, 16))
+axes = axes.flatten()
+
+for idx, (model_name, metrics) in enumerate(results.items()):
+    disp = ConfusionMatrixDisplay(metrics["cm"], display_labels=["Normal", "Attack"])
+    disp.plot(ax=axes[idx], cmap="Blues", colorbar=False)
+    axes[idx].set_title(f"{model_name}\nAccuracy = {metrics['acc']:.3f}")
+
+for i in range(idx + 1, len(axes)):
+    axes[i].axis("off")
+
+plt.tight_layout()
+plt.show()
+
+# ===============================================
+# 9. SAVE RESULTS TO TEXT FILE
+# ===============================================
+summary_file = "updated_model_results.txt"
+
+with open(summary_file, "w") as f:
+    f.write("UPDATED MODEL PERFORMANCE SUMMARY\n")
+    f.write("=" * 65 + "\n\n")
+
+    for model_name, info in results.items():
+        rep = info["report"]
+        cm = info["cm"]
+
+        f.write(f"MODEL: {model_name}\n")
+        f.write("-" * 50 + "\n")
+        f.write(f"Accuracy       : {info['acc']:.4f}\n")
+        f.write(f"Precision (Atk): {rep['1']['precision']:.4f}\n")
+        f.write(f"Recall (Atk)   : {rep['1']['recall']:.4f}\n")
+        f.write(f"F1 Score (Atk) : {rep['1']['f1-score']:.4f}\n\n")
+
+        f.write("Confusion Matrix:\n")
+        f.write(f"TN = {cm[0,0]}, FP = {cm[0,1]}\n")
+        f.write(f"FN = {cm[1,0]}, TP = {cm[1,1]}\n")
+        f.write("\n" + "-" * 50 + "\n")
+
+print(f"\nSummary file saved as → {summary_file}")

network_attack_detection.py

@@ -0,0 +1,109 @@
+# 1. IMPORTS
+import pandas as pd
+import numpy as np
+
+from sklearn.preprocessing import LabelEncoder, StandardScaler
+from sklearn.model_selection import train_test_split
+from sklearn.decomposition import PCA
+from imblearn.over_sampling import SMOTE
+
+from sklearn.linear_model import LogisticRegression
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.naive_bayes import GaussianNB
+from sklearn.svm import SVC
+from xgboost import XGBClassifier
+
+from sklearn.metrics import accuracy_score, classification_report, confusion_matrix
+
+
+# 2. LOAD DATA
+df = pd.read_csv("Friday-WorkingHours-Afternoon-DDos.pcap_ISCX.csv", low_memory=False)
+df.columns = df.columns.str.strip()
+print("Loaded:", df.shape)
+
+
+# 3. REMOVE NON-FEATURE COLUMNS
+remove_list = ["Flow ID", "Timestamp", "Source IP", "Destination IP"]
+df = df.drop([c for c in remove_list if c in df.columns], axis=1)
+
+
+# 4. CLEAN DATA
+df = df.replace([np.inf, -np.inf], np.nan)
+df = df.dropna()
+print("Cleaned:", df.shape)
+
+
+# 5. ENCODE LABEL
+enc = LabelEncoder()
+df["Attack"] = enc.fit_transform(df["Label"])
+df = df.drop(columns=["Label"])
+print("Classes:", enc.classes_)
+
+
+# 6. SPLIT FEATURES/TARGET
+X = df.drop(columns=["Attack"])
+y = df["Attack"]
+
+
+# 7. SMOTE BALANCING
+sm = SMOTE(random_state=11)
+X_bal, y_bal = sm.fit_resample(X, y)
+print("Balanced:", X_bal.shape)
+
+
+# 8. SCALING
+sc = StandardScaler()
+X_norm = sc.fit_transform(X_bal)
+
+
+# 9. PCA REDUCTION
+pca = PCA(n_components=30, random_state=11)
+X_pca = pca.fit_transform(X_norm)
+
+X_train, X_test, y_train, y_test = train_test_split(
+    X_pca, y_bal, test_size=0.25, random_state=11, stratify=y_bal
+)
+
+
+# 10. DEFINE MODELS
+models = {
+    "LogReg": LogisticRegression(max_iter=2000),
+    "DecisionTree": DecisionTreeClassifier(max_depth=20),
+    "RandomForest": RandomForestClassifier(n_estimators=350),
+    "NaiveBayes": GaussianNB(),
+    "SVM": SVC(kernel="rbf", gamma="scale"),
+    "XGBoost": XGBClassifier(
+        n_estimators=350,
+        max_depth=7,
+        learning_rate=0.07,
+        subsample=0.85,
+        colsample_bytree=0.85,
+        eval_metric="logloss"
+    )
+}
+
+
+# 11. TRAIN & EVALUATE
+scores = {}
+
+for name, clf in models.items():
+    print(f"\nMODEL: {name}")
+    clf.fit(X_train, y_train)
+    preds = clf.predict(X_test)
+
+    acc = accuracy_score(y_test, preds)
+    scores[name] = acc * 100
+
+    print("Accuracy:", round(acc * 100, 3), "%")
+    print(classification_report(y_test, preds))
+
+
+# 12. FINAL RESULT TABLE
+results = (
+    pd.DataFrame(scores.items(), columns=["Model", "Accuracy (%)"])
+      .sort_values(by="Accuracy (%)", ascending=False)
+)
+
+print("\nFINAL MODEL SCORES")
+print(results)

preprocessing.py

@@ -0,0 +1,104 @@
+# ================================
+# KDD Dataset Preprocessing Script
+# ================================
+
+import pandas as pd
+import numpy as np
+from sklearn.preprocessing import StandardScaler, LabelEncoder
+from sklearn.impute import SimpleImputer
+from imblearn.over_sampling import SMOTE
+from collections import Counter
+
+# ================================
+# Load Train and Test Data
+# ================================
+train = pd.read_csv('kdd_train.csv')
+test = pd.read_csv('kdd_test.csv')
+
+print("Train Shape:", train.shape)
+print("Test Shape:", test.shape)
+
+# ================================
+# Split Features and Labels
+# ================================
+X_train = train.drop('labels', axis=1)
+y_train = train['labels']
+X_test = test.drop('labels', axis=1)
+y_test = test['labels']
+
+# ================================
+# Identify Numeric and Categorical Columns
+# ================================
+num_cols = X_train.select_dtypes(include=['int64', 'float64']).columns
+cat_cols = X_train.select_dtypes(include=['object']).columns
+
+# ================================
+# Handle Missing Values
+# ================================
+
+# For numeric columns → fill missing values with mean
+num_imputer = SimpleImputer(strategy='mean')
+X_train[num_cols] = num_imputer.fit_transform(X_train[num_cols])
+X_test[num_cols] = num_imputer.transform(X_test[num_cols])
+
+# For categorical columns → fill missing with 'missing'
+X_train[cat_cols] = X_train[cat_cols].fillna('missing')
+X_test[cat_cols] = X_test[cat_cols].fillna('missing')
+
+# ================================
+# Encode Categorical Columns
+# ================================
+for col in cat_cols:
+    le = LabelEncoder()
+    all_values = pd.concat([X_train[col], X_test[col]], axis=0).astype(str)
+    le.fit(all_values)
+    X_train[col] = le.transform(X_train[col].astype(str))
+    X_test[col] = le.transform(X_test[col].astype(str))
+
+# ================================
+# Scale Numeric Features
+# ================================
+scaler = StandardScaler()
+X_train_scaled = scaler.fit_transform(X_train)
+X_test_scaled = scaler.transform(X_test)
+
+# ================================
+# Handle Class Imbalance with SMOTE
+# ================================
+print("Class distribution before SMOTE:", Counter(y_train))
+
+# Remove rare classes (with only 1 sample)
+class_counts = y_train.value_counts()
+valid_classes = class_counts[class_counts > 1].index
+mask = y_train.isin(valid_classes)
+X_train_filtered = X_train_scaled[mask]
+y_train_filtered = y_train[mask]
+
+# Apply SMOTE to balance dataset
+smote = SMOTE(random_state=42, k_neighbors=1)
+X_train_balanced, y_train_balanced = smote.fit_resample(X_train_filtered, y_train_filtered)
+
+print("Class distribution after SMOTE:", Counter(y_train_balanced))
+
+# ================================
+# Convert Back to DataFrames
+# ================================
+X_train_final = pd.DataFrame(X_train_balanced)
+y_train_final = pd.DataFrame(y_train_balanced, columns=['labels'])
+X_test_final = pd.DataFrame(X_test_scaled)
+y_test_final = pd.DataFrame(y_test, columns=['labels'])
+
+# ================================
+# Save Preprocessed Data
+# ================================
+X_train_final.to_csv("X_train_preprocessed.csv", index=False)
+y_train_final.to_csv("y_train_preprocessed.csv", index=False)
+X_test_final.to_csv("X_test_preprocessed.csv", index=False)
+y_test_final.to_csv("y_test_preprocessed.csv", index=False)
+
+print("\n Preprocessing complete!")
+print("Files saved:")
+print(" - X_train_preprocessed.csv")
+print(" - y_train_preprocessed.csv")
+print(" - X_test_preprocessed.csv")
+print(" - y_test_preprocessed.csv")