Feature extraction and model training

app.py

@@ -3,14 +3,13 @@ import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 import seaborn as sns
+from collections import Counter
 from sklearn.model_selection import train_test_split
 from sklearn.preprocessing import StandardScaler, LabelEncoder
 from sklearn.feature_selection import SelectKBest, f_classif
 from sklearn.impute import SimpleImputer
-from imblearn.over_sampling import SMOTE
+from imblearn.over_sampling import SMOTE, RandomOverSampler
 from sklearn.metrics import accuracy_score, classification_report, mean_squared_error, mean_absolute_error, r2_score
-
-# Import ML Models
 from sklearn.neighbors import KNeighborsClassifier, KNeighborsRegressor
 from sklearn.svm import SVC, SVR
 from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor
@@ -49,20 +48,9 @@ problems = {
 
 problem = st.sidebar.selectbox("Choose a Problem:", problems[task][model])
 
-dataset_mapping = {name: f"datasets/{name.lower().replace(' ', '_')}.csv" for sublist in problems.values() for model in sublist for name in sublist[model]}
-
-# # Dataset Selection (User selects a pre-existing fake dataset)
-# dataset_mapping = {
-#     "Spam Detection": "datasets/spam_detection.csv",
-#     "Disease Prediction": "datasets/disease_prediction.csv",
-#     "Image Recognition": "datasets/image_recognition.csv",
-#     "Text Classification": "datasets/text_classification.csv",
-#     "Fraud Detection": "datasets/fraud_detection.csv",
-#     "Customer Segmentation": "datasets/customer_segmentation.csv",
-#     "Loan Approval": "datasets/loan_approval.csv",
-#     "House Price Prediction": "datasets/house_price_prediction.csv",
-#     "Sales Forecasting": "datasets/sales_forecasting.csv",
-# }
+# Dataset Mapping (Dynamic)
+dataset_mapping = {name: f"datasets/{name.lower().replace(' ', '_')}.csv" 
+                   for sublist in problems.values() for model in sublist for name in sublist[model]}
 
 dataset_path = dataset_mapping.get(problem, "datasets/spam_detection.csv")
 df = pd.read_csv(dataset_path)
@@ -92,8 +80,8 @@ for col in df.select_dtypes(include=['object']).columns:
     df[col] = label_encoders[col].fit_transform(df[col])
 
 # Split Data
-X = df.iloc[:, :-1]  # Features
-y = df.iloc[:, -1]   # Target
+X = df.iloc[:, :-1]
+y = df.iloc[:, -1]
 X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
 
 # Feature Scaling
@@ -102,35 +90,31 @@ X_train = scaler.fit_transform(X_train)
 X_test = scaler.transform(X_test)
 
 # Feature Selection
-selector = SelectKBest(score_func=f_classif, k=min(5, X.shape[1]))  # Ensure k does not exceed available features
+selector = SelectKBest(score_func=f_classif, k=min(X.shape[1], 5))
 X_train = selector.fit_transform(X_train, y_train)
 X_test = selector.transform(X_test)
 
 # Handle imbalanced data
 if task == "Classification":
-    smote = SMOTE()
-    X_train, y_train = smote.fit_resample(X_train, y_train)
+    class_counts = Counter(y_train)
+    min_class_samples = min(class_counts.values())
 
-# Model Initialization
-if task == "Classification":
-    n_neighbors = min(5, len(y_train))  # Ensure k is valid
-    model_mapping = {
-        "KNN": KNeighborsClassifier(n_neighbors=n_neighbors),
-        "SVM": SVC(),
-        "Random Forest": RandomForestClassifier(),
-        "Decision Tree": DecisionTreeClassifier(),
-        "Perceptron": Perceptron()
-    }
-else:
-    n_neighbors = min(5, len(y_train))  # Ensure k is valid
-    model_mapping = {
-        "KNN": KNeighborsRegressor(n_neighbors=n_neighbors),
-        "SVM": SVR(),
-        "Random Forest": RandomForestRegressor(),
-        "Decision Tree": DecisionTreeRegressor(),
-        "Perceptron": Perceptron()
-    }
+    if min_class_samples > 5:
+        smote = SMOTE()
+        X_train, y_train = smote.fit_resample(X_train, y_train)
+    else:
+        ros = RandomOverSampler()
+        X_train, y_train = ros.fit_resample(X_train, y_train)
 
+# Model Initialization
+n_neighbors = min(5, len(y_train))
+model_mapping = {
+    "KNN": KNeighborsClassifier(n_neighbors=n_neighbors) if task == "Classification" else KNeighborsRegressor(n_neighbors=n_neighbors),
+    "SVM": SVC() if task == "Classification" else SVR(),
+    "Random Forest": RandomForestClassifier() if task == "Classification" else RandomForestRegressor(),
+    "Decision Tree": DecisionTreeClassifier() if task == "Classification" else DecisionTreeRegressor(),
+    "Perceptron": Perceptron()
+}
 model_instance = model_mapping[model]
 
 # Train Model
@@ -139,52 +123,30 @@ y_pred = model_instance.predict(X_test)
 
 # Model Evaluation
 st.subheader("📊 Model Evaluation")
-
 if task == "Classification":
     accuracy = accuracy_score(y_test, y_pred)
     report = classification_report(y_test, y_pred, output_dict=True)
-
     st.write(f"**Accuracy:** {accuracy:.2f}")
-    st.json(report)  # Shows detailed structured metrics
-    
-elif task == "Regression":
+    st.json(report)
+else:
     mse = mean_squared_error(y_test, y_pred)
     mae = mean_absolute_error(y_test, y_pred)
     r2 = r2_score(y_test, y_pred)
-    
-    st.write(f"**Mean Squared Error (MSE):** {mse:.4f}")
-    st.write(f"**Mean Absolute Error (MAE):** {mae:.4f}")
-    st.write(f"**R² Score:** {r2:.4f}")
+    st.write(f"**MSE:** {mse:.4f}, **MAE:** {mae:.4f}, **R² Score:** {r2:.4f}")
 
 # Data Visualization
 st.subheader("📈 Data Visualization")
 
-# Heatmap
 st.write("### 🔥 Feature Correlation")
-plt.figure(figsize=(8, 5))
-sns.heatmap(df.corr(), annot=True, cmap="coolwarm")
-st.pyplot(plt)
-
-# Pair Plot
-st.write("### 📊 Pair Plot of Features")
-sns.pairplot(df, diag_kind='kde')
-st.pyplot()
-
-# Feature Importance (for tree-based models)
-if model in ["Random Forest", "Decision Tree"]:
-    feature_importances = model_instance.feature_importances_
-    feature_names = X.columns
-    importance_df = pd.DataFrame({"Feature": feature_names, "Importance": feature_importances})
-    importance_df = importance_df.sort_values(by="Importance", ascending=False)
-    
+fig, ax = plt.subplots(figsize=(8, 5))
+sns.heatmap(df.corr(), annot=True, cmap="coolwarm", ax=ax)
+st.pyplot(fig)
+
+if model in ["Random Forest", "Decision Tree"] and hasattr(model_instance, "feature_importances_"):
+    importance_df = pd.DataFrame({"Feature": X.columns, "Importance": model_instance.feature_importances_}).sort_values(by="Importance", ascending=False)
     st.write("### 🌟 Feature Importance")
     fig, ax = plt.subplots()
     sns.barplot(x=importance_df["Importance"], y=importance_df["Feature"], ax=ax)
     st.pyplot(fig)
 
-# Download Code
-st.download_button("🐍 Download Python Code (.py)", "ai_model.py")
-st.download_button("📓 Download Notebook (.ipynb)", "ai_model.ipynb")
-st.markdown("[🚀 Open in Colab](https://colab.research.google.com/)")
-
-st.success("Code generated! Download and do magic! ✨")
+st.success("Code generated! 🚀 Download & run!")