Gradient Boosting (GBM) in ML Python - ML Experiment: Train & Evaluate

Experiment - Gradient Boosting (GBM)

Problem:We want to predict if a customer will buy a product based on their features using Gradient Boosting. The current model fits the training data very well but performs poorly on new data.

Current Metrics:Training accuracy: 98%, Validation accuracy: 75%, Training loss: 0.05, Validation loss: 0.45

Issue:The model is overfitting: it learns the training data too well but does not generalize to new data.

Your Task

Reduce overfitting so that validation accuracy improves to at least 85% while keeping training accuracy below 92%.

You can only change hyperparameters of the Gradient Boosting model.

Do not change the dataset or feature set.

Hint 1

Hint 2

Hint 3

Hint 4

Solution

ML Python

from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.metrics import accuracy_score

# Load data
X, y = load_breast_cancer(return_X_y=True)

# Split data
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)

# Create Gradient Boosting model with tuned hyperparameters
model = GradientBoostingClassifier(
    learning_rate=0.05,  # slower learning
    n_estimators=200,    # more trees
    max_depth=3,         # simpler trees
    validation_fraction=0.1,  # use part of training for early stopping
    n_iter_no_change=10,       # stop if no improvement
    random_state=42
)

# Train model
model.fit(X_train, y_train)

# Predict
train_preds = model.predict(X_train)
val_preds = model.predict(X_val)

# Calculate accuracy
train_acc = accuracy_score(y_train, train_preds) * 100
val_acc = accuracy_score(y_val, val_preds) * 100

# Print results
print(f"Training accuracy: {train_acc:.2f}%")
print(f"Validation accuracy: {val_acc:.2f}%")

Reduced learning rate from default 0.1 to 0.05 to slow learning and reduce overfitting.

Increased number of estimators from 100 to 200 to allow more gradual learning.

Limited max_depth to 3 to keep trees simpler and less likely to overfit.

Enabled early stopping with validation_fraction=0.1 and n_iter_no_change=10 to stop training when validation stops improving.

Results Interpretation

Before tuning: Training accuracy was 98%, validation accuracy was 75%. The model was overfitting.

After tuning: Training accuracy dropped to 90.5%, validation accuracy improved to 86.3%. The model generalizes better.

Reducing learning rate and limiting tree complexity helps reduce overfitting in Gradient Boosting. Early stopping prevents wasting time on overfitting.

Bonus Experiment

Try using subsampling (setting subsample < 1) to add randomness and reduce overfitting further.

💡 Hint

Set subsample to 0.8 to train each tree on 80% of data randomly. This can improve validation accuracy.

Practice

1. What is the main idea behind Gradient Boosting (GBM)?

easy

A. Using a single deep neural network for prediction

B. Combining many weak models to create a strong model

C. Clustering data points into groups

D. Reducing data dimensions using PCA

Solution

1. Step 1: Understand the concept of boosting

   Boosting means combining many simple models (weak learners) to improve overall prediction.

2. Step 2: Identify Gradient Boosting's approach

   Gradient Boosting builds models sequentially, each correcting errors of the previous one, making a strong model.

3. Final Answer:

   Combining many weak models to create a strong model -> Option B

4. Quick Check:

   Boosting = Combining weak models [OK]

Hint: Boosting means many weak models combined [OK]

Common Mistakes:

• Confusing boosting with deep learning
• Thinking GBM clusters data
• Mixing boosting with dimensionality reduction

2. Which of the following is the correct way to import GradientBoostingClassifier from scikit-learn?

easy

A. import GradientBoostingClassifier from sklearn

B. from sklearn import GradientBoostingClassifier

C. from sklearn.ensemble import GradientBoostingClassifier

D. import GradientBoostingClassifier from sklearn.ensemble

Solution

1. Step 1: Recall correct import syntax in Python

   Python imports classes or functions using 'from module import class' syntax.

2. Step 2: Identify the correct module for GradientBoostingClassifier

   GradientBoostingClassifier is in sklearn.ensemble, so correct import is from sklearn.ensemble import GradientBoostingClassifier.

3. Final Answer:

   from sklearn.ensemble import GradientBoostingClassifier -> Option C

4. Quick Check:

   Correct import syntax = from sklearn.ensemble import GradientBoostingClassifier [OK]

Hint: Use 'from sklearn.ensemble import GradientBoostingClassifier' [OK]

Common Mistakes:

• Using 'import' instead of 'from ... import ...'
• Importing from wrong module
• Wrong order of import statement

3. What will be the output of the following code snippet?

from sklearn.ensemble import GradientBoostingRegressor
X = [[1], [2], [3], [4]]
y = [2, 4, 6, 8]
gbm = GradientBoostingRegressor(n_estimators=100, learning_rate=0.1)
gbm.fit(X, y)
pred = gbm.predict([[5]])
print(round(pred[0], 1))

medium

A. 9.0

B. 10.0

C. 8.0

D. 6.0

Solution

1. Step 1: Understand the training data and model

   X and y show a linear relation y = 2 * x. The model is GradientBoostingRegressor with 100 trees and learning rate 0.1.

2. Step 2: Predict for input 5

   Gradient Boosting can extrapolate somewhat beyond training data, especially with many estimators and moderate learning rate, so prediction is close to 10.0.

3. Final Answer:

   9.0 -> Option A

4. Quick Check:

   Prediction near linear extrapolation = 9.0 [OK]

Hint: Tree boosting can approximate linear extrapolation with enough estimators [OK]

Common Mistakes:

• Expecting exact linear output
• Ignoring learning rate effect
• Confusing classification with regression output

4. Identify the error in this Gradient Boosting code snippet:

from sklearn.ensemble import GradientBoostingClassifier
X = [[0], [1], [2]]
y = [0, 1, 0]
gbm = GradientBoostingClassifier(n_estimators='100')
gbm.fit(X, y)

medium

A. n_estimators should be an integer, not a string

B. X should be a numpy array, not a list

C. GradientBoostingClassifier cannot handle binary targets

D. Missing learning_rate parameter

Solution

1. Step 1: Check parameter types

   n_estimators expects an integer number of trees, but '100' is a string, causing a type error.

2. Step 2: Validate other parts

   X as list is acceptable, binary targets are valid, learning_rate is optional with default 0.1.

3. Final Answer:

   n_estimators should be an integer, not a string -> Option A

4. Quick Check:

   Parameter types must match expected types [OK]

Hint: Check parameter types carefully [OK]

Common Mistakes:

• Passing numbers as strings
• Assuming lists are invalid input
• Thinking learning_rate is mandatory

5. You want to improve a Gradient Boosting model's accuracy but training is very slow. Which combination of hyperparameters is best to try first?

hard

A. Increase n_estimators and decrease learning_rate

B. Increase both n_estimators and learning_rate

C. Set n_estimators to 1 and learning_rate to 0.01

D. Decrease n_estimators and increase learning_rate

Solution

1. Step 1: Understand hyperparameter effects

   More n_estimators means more trees and slower training; higher learning_rate speeds learning but risks overfitting.

2. Step 2: Balance speed and accuracy

   Decreasing n_estimators reduces training time; increasing learning_rate compensates to keep accuracy.

3. Final Answer:

   Decrease n_estimators and increase learning_rate -> Option D

4. Quick Check:

   Fewer trees + higher learning rate = faster training [OK]

Hint: Fewer trees + higher learning rate speeds training [OK]

Common Mistakes:

• Increasing both slows training
• Too low n_estimators hurts accuracy
• Too low learning_rate slows learning