Imbalanced class handling (SMOTE, class weights) in ML Python - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the SMOTE class from the imblearn library.

ML Python

from imblearn.[1] import SMOTE

Drag options to blanks, or click blank then click option'

Aover_sampling

Bunder_sampling

Censemble

Dmetrics

Attempts:

3 left

💡 Hint

Common Mistakes

Importing SMOTE from the wrong module like under_sampling or metrics.

Misspelling the module name.

The SMOTE class is located in the over_sampling module of the imblearn library. This module contains techniques to increase minority class samples.

2fill in blank

medium

Complete the code to create a SMOTE object with a random state for reproducibility.

ML Python

smote = SMOTE(random_state=[1])

Drag options to blanks, or click blank then click option'

ANone

B42

C'seed'

D0.5

Attempts:

3 left

💡 Hint

Common Mistakes

Using a string instead of an integer for random_state.

Passing None which disables reproducibility.

Setting random_state=42 ensures the SMOTE sampling is reproducible. The value 42 is commonly used as a fixed seed.

3fill in blank

hard

Fix the error in the code to fit and resample the training data using SMOTE.

ML Python

X_resampled, y_resampled = smote.[1](X_train, y_train)

Drag options to blanks, or click blank then click option'

Afit

Btransform

Cfit_transform

Dsample

Attempts:

3 left

💡 Hint

Common Mistakes

Using fit() which returns the object but not data.

Using transform() which requires prior fitting.

The fit_transform method fits the SMOTE model and returns the resampled data. Using only fit does not return the resampled data.

4fill in blank

hard

Fill both blanks to create a logistic regression model with balanced class weights and fit it on training data.

ML Python

model = LogisticRegression(class_weight=[1])
model.[2](X_train, y_train)

Drag options to blanks, or click blank then click option'

A'balanced'

Bfit

Cpredict

DNone

Attempts:

3 left

💡 Hint

Common Mistakes

Passing None or wrong string to class_weight.

Using predict instead of fit to train the model.

Setting class_weight='balanced' tells the model to adjust weights inversely proportional to class frequencies. Then fit trains the model on data.

5fill in blank

hard

Fill all three blanks to compute class weights manually and train a logistic regression model with them.

ML Python

from sklearn.utils.class_weight import compute_class_weight

weights = compute_class_weight(class_weight=[1], classes=[2], y=[3])
class_weights = dict(zip([2], weights))
model = LogisticRegression(class_weight=class_weights)
model.fit(X_train, y_train)

Drag options to blanks, or click blank then click option'

A'balanced'

Bnp.unique(y_train)

Cy_train

DNone

Attempts:

3 left

💡 Hint

Common Mistakes

Passing None instead of 'balanced' to compute_class_weight.

Using wrong variables for classes or y parameters.

To compute class weights manually, use class_weight='balanced', pass the unique classes from training labels, and the training labels themselves. Then create a dictionary for the model.