Imbalanced class handling (SMOTE, class weights) in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Why use SMOTE for imbalanced data?

Imagine you have a dataset where one class is much smaller than the other. Why would you use SMOTE (Synthetic Minority Over-sampling Technique) instead of just duplicating minority class samples?

A SMOTE creates new synthetic samples by mixing existing minority samples, which helps the model learn better decision boundaries.

B SMOTE removes majority class samples to balance the dataset, reducing training time.

C SMOTE duplicates minority samples exactly to increase their count without changing data distribution.

D SMOTE randomly deletes samples from both classes to balance the dataset.

💡 Hint

Think about how creating new data points can help the model generalize better than just copying existing ones.

SMOTE generates new synthetic minority class samples by interpolating between existing ones. This helps the model learn a more general decision boundary rather than memorizing duplicates.

❓ Predict Output

intermediate

Output of class weight usage in logistic regression

What will be the output of the following code snippet regarding the model's class weight attribute?

ML Python

from sklearn.linear_model import LogisticRegression
model = LogisticRegression(class_weight={0:1, 1:5})
print(model.class_weight)

A {0: 1, 1: 5}

B balanced

C None

D Raises TypeError

💡 Hint

Check what value is assigned to class_weight in the model initialization.

The class_weight parameter is explicitly set to a dictionary {0:1, 1:5}, so printing model.class_weight will output this dictionary.

❓ Metrics

advanced

Choosing the best metric for imbalanced classification

You trained a model on a dataset with 95% of class 0 and 5% of class 1. Which metric is best to evaluate your model's performance on the minority class?

A Precision for class 1, to measure how many predicted positives are correct.

B Mean Squared Error, because it measures prediction error.

C Accuracy, because it shows overall correct predictions.

D Recall for class 1, to measure how many actual positives are found.

💡 Hint

Think about which metric helps find most of the minority class samples.

Recall for the minority class measures how many actual positive samples are correctly identified, which is crucial in imbalanced data to avoid missing minority cases.

🔧 Debug

advanced

Why does this SMOTE code raise an error?

What error will this code raise and why? from imblearn.over_sampling import SMOTE X = [[1,2],[3,4],[5,6]] y = [0,0,1] smote = SMOTE(sampling_strategy='minority') X_res, y_res = smote.fit_resample(X, y)

A ValueError: Expected 2D array, got 1D array instead.

B No error, code runs successfully.

C ValueError: At least 6 samples are needed to perform SMOTE.

D TypeError: 'list' object has no attribute 'fit'.

💡 Hint

SMOTE needs enough samples in the minority class to create synthetic samples.

SMOTE requires at least k_neighbors+1 samples in the minority class to generate synthetic samples. Here, only one minority sample exists, so it raises a ValueError.

❓ Model Choice

expert

Best approach for highly imbalanced multi-class classification

You have a multi-class dataset with 4 classes, where one class is only 1% of data. You want to improve model performance on the rare class. Which approach is best?

A Apply SMOTE oversampling to all classes equally before training.

B Use class weights in the model to give higher importance to the rare class.

C Remove the rare class samples to simplify the problem.

D Use accuracy as the only metric to evaluate the model.

💡 Hint

Think about how to handle imbalance without losing data or misleading metrics.

Using class weights allows the model to pay more attention to the rare class during training without changing data distribution or removing samples.