Boosting concept in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

What is the main idea behind boosting in machine learning?

Boosting is a technique used to improve model performance. What best describes its main idea?

A Using a single model and tuning its hyperparameters to reduce error.

B Training a single very deep model to capture all data patterns at once.

C Randomly selecting features to train multiple independent models and averaging their results.

D Combining many weak models sequentially, where each new model focuses on correcting errors of the previous ones.

💡 Hint

Think about how boosting tries to fix mistakes step by step.

Boosting builds models one after another. Each new model tries to fix the errors made by the models before it, improving overall accuracy.

❓ Model Choice

intermediate

Which model is commonly used as a weak learner in boosting algorithms?

Boosting algorithms often use a simple model repeatedly. Which of these is typically used?

A K-nearest neighbors with large k

B Deep convolutional neural networks

C Decision stumps (trees with one split)

D Support vector machines with complex kernels

💡 Hint

Weak learners are simple and fast to train.

Decision stumps are simple trees with one split, making them weak learners ideal for boosting to combine and improve.

❓ Metrics

advanced

In AdaBoost, how are the weights of training samples updated after each iteration?

AdaBoost adjusts sample weights to focus on hard examples. How exactly are these weights updated?

A Weights of all samples are kept the same throughout training.

B Weights of misclassified samples are increased, and weights of correctly classified samples are decreased.

C Weights of correctly classified samples are increased, and misclassified samples are decreased.

D Weights are randomly shuffled after each iteration.

💡 Hint

Think about how boosting focuses on mistakes.

AdaBoost increases weights for samples the model got wrong, so the next model pays more attention to them.

🔧 Debug

advanced

What error will this Python snippet raise when implementing a simple boosting weight update?

Consider this code snippet for updating sample weights in boosting: weights = [0.1, 0.2, 0.3, 0.4] predictions = [1, 0, 1, 0] labels = [1, 1, 0, 0] for i in range(len(weights)): if predictions[i] != labels[i] weights[i] *= 1.5 else: weights[i] *= 0.5 print(weights)

ML Python

weights = [0.1, 0.2, 0.3, 0.4]
predictions = [1, 0, 1, 0]
labels = [1, 1, 0, 0]
for i in range(len(weights)):
    if predictions[i] != labels[i]:
        weights[i] *= 1.5
    else:
        weights[i] *= 0.5
print(weights)

A SyntaxError due to missing colon after the if statement

B TypeError because weights list cannot be multiplied by float

C IndexError because of out-of-range index access

D No error, prints updated weights

💡 Hint

Check the syntax of the if statement line.

The if statement is missing a colon at the end, causing a SyntaxError.

❓ Predict Output

expert

What is the output of this simplified AdaBoost weight update code?

Given the code below, what is the final list of weights printed? weights = [0.25, 0.25, 0.25, 0.25] predictions = [1, 0, 1, 0] labels = [1, 1, 0, 0] alpha = 0.7 for i in range(len(weights)): if predictions[i] != labels[i]: weights[i] *= (2.718 ** alpha) else: weights[i] *= (2.718 ** (-alpha)) print([round(w, 3) for w in weights])

ML Python

weights = [0.25, 0.25, 0.25, 0.25]
predictions = [1, 0, 1, 0]
labels = [1, 1, 0, 0]
alpha = 0.7
for i in range(len(weights)):
    if predictions[i] != labels[i]:
        weights[i] *= (2.718 ** alpha)
    else:
        weights[i] *= (2.718 ** (-alpha))
print([round(w, 3) for w in weights])

A [0.125, 0.5, 0.5, 0.125]

B [0.125, 0.5, 0.125, 0.5]

C [0.5, 0.125, 0.125, 0.5]

D [0.5, 0.125, 0.5, 0.125]

💡 Hint

Calculate e^0.7 ≈ 2.013 and e^-0.7 ≈ 0.497, then multiply accordingly.

For each index: - If prediction != label, multiply weight by ~2.013 - Else multiply by ~0.497 Weights start at 0.25. Index 0: pred=1, label=1 (equal) → 0.25*0.497=0.12425 Index 1: pred=0, label=1 (not equal) → 0.25*2.013=0.50325 Index 2: pred=1, label=0 (not equal) → 0.25*2.013=0.50325 Index 3: pred=0, label=0 (equal) → 0.25*0.497=0.12425 Rounded: [0.124, 0.503, 0.503, 0.124] Option A matches closest with [0.125, 0.5, 0.5, 0.125].