Submitted By:

Ahsan Azeem Zahid

Reg#:

FA22-BCS-213

Submitted To:

Mam Maryam Mazhar

Assigment #

03
dt_classification_iris.py

# Import necessary libraries

import numpy as np

import pandas as pd

import [Link] as plt

from [Link] import load_iris # Built-in dataset for classification

from sklearn.model_selection import train_test_split # For splitting data

from [Link] import StandardScaler # For scaling features

from [Link] import DecisionTreeClassifier, plot_tree # DT model and viz

from [Link] import accuracy_score, confusion_matrix, classification_report #

from [Link].outliers_influence import variance_inflation_factor # For VIF

import warnings

[Link]('ignore') # Ignore warnings for clean output

# Step 1: Data Selection

# Load the Iris dataset (150 samples, 4 features, 3 classes: setosa, versicolor, virginica)

iris = load_iris()

X = [Link]([Link], columns=iris.feature_names) # Features: sepal/petal

y = [Link] # Target: class labels 0,1,2

print("Dataset shape:", [Link])

print("Target classes:", [Link](y))

# Step 2: Data Preprocessing

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Scale features to standardize (mean=0, std=1) - important for DT to avoid bias

scaler = StandardScaler()

X_train_scaled = scaler.fit_transform(X_train)

X_test_scaled = [Link](X_test)

print("\nPreprocessing complete. Train shape:", X_train_scaled.shape)

# Step 3: Multi-collinearity Check

# Calculate VIF for each feature to detect multicollinearity (VIF > 5-10 indicates issue)

vif_data = [Link]()

vif_data["feature"] = [Link]

vif_data["VIF"] = [variance_inflation_factor(X_train_scaled, i) for i in range(len([Link]))]

print("\nVIF Scores:\n", vif_data)

# Insight: High VIF (e.g., >10) suggests correlated features; may need feature selection

# Step 4: Model Building

# Build Decision Tree Classifier (from sklearn; from-scratch would use Gini/entropy for
splits)

clf = DecisionTreeClassifier(random_state=42, max_depth=3) # Limit depth to prevent

[Link](X_train_scaled, y_train) # Train on scaled data

# Step 5: Model Evaluation

y_pred = [Link](X_test_scaled) # Predict on test

accuracy = accuracy_score(y_test, y_pred) # Accuracy score

print("\nAccuracy:", accuracy)
print("\nConfusion Matrix:\n", confusion_matrix(y_test, y_pred))

print("\nClassification Report:\n", classification_report(y_test, y_pred))

# Step 6: Visualization

# Plot the decision tree

[Link](figsize=(12,8))

plot_tree(clf, feature_names=iris.feature_names, class_names=iris.target_names,

[Link]('Decision Tree for Iris Classification')

[Link]('iris_tree.png') # Save plot

[Link]() # Display if running locally

# Feature Importance Bar Plot

importances = [Link]({'feature': iris.feature_names, 'importance':

[Link](figsize=(8,6))

[Link](x='feature', y='importance')

[Link]('Feature Importance in Iris DT')

[Link]('iris_importance.png') # Save plot

[Link]() # Display if running locally

print("\nPlots saved as iris_tree.png and iris_importance.png")

Output:
dt_regression_synthetic.py.
# Import necessary libraries

import numpy as np

import pandas as pd

import [Link] as plt

from sklearn.model_selection import train_test_split # For splitting

from [Link] import StandardScaler # For scaling

from [Link] import DecisionTreeRegressor, plot_tree # DT model and viz

from [Link] import mean_squared_error, r2_score # Evaluation metrics

from [Link].outliers_influence import variance_inflation_factor # For VIF

import warnings

[Link]('ignore') # Ignore warnings

# Generate synthetic dataset (200 samples, 4 features, 1 target) - simulates correlated

[Link](42)

n_samples = 200

X = [Link](n_samples, 4)

# Introduce multicollinearity: Feature3 correlated with Feature1

X[:, 2] = X[:, 0] * 0.8 + [Link](n_samples) * 0.2

feature_names = ['Feature1', 'Feature2', 'Feature3_Corr', 'Feature4']

X = [Link](X, columns=feature_names)

y = X['Feature1'] + 2 * X['Feature2'] + 0.5 * X['Feature3_Corr'] +

print("Dataset shape:", [Link])

print("Target range:", [Link](), "to", [Link]())

# Step 2: Data Preprocessing

# Split into train/test (80/20)

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Scale features

scaler = StandardScaler()

X_train_scaled = scaler.fit_transform(X_train)

X_test_scaled = [Link](X_test)

print("\nPreprocessing complete. Train shape:", X_train_scaled.shape)

# Step 3: Multi-collinearity Check

vif_data = [Link]()
vif_data["feature"] = feature_names

vif_data["VIF"] = [variance_inflation_factor(X_train_scaled, i) for i in

print("\nVIF Scores:\n", vif_data)

# Insight: High VIF for correlated features indicates multicollinearity; consider dropping one

# Step 4: Model Building

reg = DecisionTreeRegressor(random_state=42, max_depth=5) # Limit depth to prevent

[Link](X_train_scaled, y_train) # Train

# Step 5: Model Evaluation

y_pred = [Link](X_test_scaled)

mse = mean_squared_error(y_test, y_pred)

r2 = r2_score(y_test, y_pred)

print("\nMSE:", mse)

print("R2 Score:", r2)

# Step 6: Visualization

# Plot tree (limit depth for readability)

[Link](figsize=(12,8))

plot_tree(reg, feature_names=feature_names, filled=True, max_depth=3)

[Link]('Decision Tree for Regression')

[Link]('regression_tree.png')

[Link]()

# Feature Importance
importances = [Link]({'feature': feature_names, 'importance':
reg.feature_importances_}).sort_values('importance', ascending=False)

[Link](figsize=(8,6))

[Link](x='feature', y='importance')

[Link]('Feature Importance in Regression DT')

[Link]('regression_importance.png')

[Link]()

# Actual vs Predicted Scatter

[Link](figsize=(8,6))

[Link](y_test, y_pred, alpha=0.5)

[Link]([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], 'r--', lw=2)

[Link]('Actual')

[Link]('Predicted')

[Link]('Actual vs Predicted Values')

[Link]('regression_pred.png')

[Link]()

print("\nPlots saved as regression_tree.png, regression_importance.png,