Semi-supervised learning basics in ML Python - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the semi-supervised learning module from scikit-learn.

ML Python

from sklearn import [1]

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

Asemi_supervised

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'semi-supervised' with a dash instead of underscore.

Trying to import from 'supervised' or 'unsupervised' instead.

The semi-supervised learning module in scikit-learn is called 'semi_supervised'.

2fill in blank

medium

Complete the code to create a LabelPropagation model instance.

ML Python

model = [1]()

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

ALabelPropagation

BLabelSpreading

CKNeighborsClassifier

DRandomForestClassifier

Attempts:

3 left

💡 Hint

Common Mistakes

Using supervised classifiers like KNeighborsClassifier or RandomForestClassifier.

Confusing LabelPropagation with LabelSpreading.

LabelPropagation is a common semi-supervised learning model in scikit-learn.

3fill in blank

hard

Fix the error in the code to fit the semi-supervised model with data X and labels y, where unlabeled points are marked as -1.

ML Python

model.fit([1], y)

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

Amodel

By

CX

Dunlabeled

Attempts:

3 left

💡 Hint

Common Mistakes

Passing labels y as the first argument instead of X.

Passing the model itself or an undefined variable.

The fit method requires the feature data X as the first argument.

4fill in blank

hard

Fill both blanks to create a dictionary of predicted labels for unlabeled data points.

ML Python

predictions = {i: model.[1]()[i] for i in range(len([2])) if y[i] == -1}

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

Apredict

BX

Cfit

Dscore

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'fit' or 'score' instead of 'predict'.

Iterating over y instead of X.

We use model.predict() to get predicted labels and iterate over X to find unlabeled points.

5fill in blank

hard

Fill all three blanks to create a semi-supervised learning pipeline with LabelSpreading, fit it, and predict labels.

ML Python

from sklearn.semi_supervised import [1]
model = [2](kernel=[3])
model.fit(X, y)
predicted_labels = model.predict(X)

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

ALabelSpreading

B'rbf'

C'linear'

DLabelPropagation

Attempts:

3 left

💡 Hint

Common Mistakes

Using LabelPropagation instead of LabelSpreading.

Using 'linear' kernel which is less common here.

LabelSpreading is imported and instantiated with kernel='rbf' for smooth label spreading.