Collaborative filtering in ML Python - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a user-item rating matrix using pandas.

ML Python

import pandas as pd
ratings = pd.DataFrame({
    'user_id': [1, 2, 1, 3],
    'item_id': [101, 101, 102, 103],
    'rating': [5, 4, 3, 2]
})
rating_matrix = ratings.pivot(index='user_id', columns=[1], values='rating')

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

A'user_id'

B'rating'

C'score'

D'item_id'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'rating' instead of 'item_id' for columns in pivot.

Using 'user_id' for columns which is already the index.

The pivot function needs the column name that represents items to create the user-item matrix. Here, 'item_id' is the correct column.

2fill in blank

medium

Complete the code to calculate cosine similarity between users using sklearn.

ML Python

from sklearn.metrics.pairwise import [1]
import numpy as np
user_ratings = np.array([[5, 3, 0], [4, 0, 0], [1, 1, 0]])
similarity = [1](user_ratings)

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

Apairwise_distances

Beuclidean_distances

Ccosine_similarity

Dmanhattan_distances

Attempts:

3 left

💡 Hint

Common Mistakes

Using distance functions instead of similarity.

Not importing the correct function from sklearn.

Cosine similarity measures how similar two users are based on their rating vectors. sklearn's cosine_similarity function computes this.

3fill in blank

hard

Fix the error in the code to predict a user's rating for an item using weighted average of neighbors' ratings.

ML Python

import numpy as np
user_similarities = np.array([0.9, 0.8, 0.4])
neighbor_ratings = np.array([5, 3, 4])
predicted_rating = np.sum(user_similarities * [1]) / np.sum(user_similarities)

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

Aneighbor_ratings

Buser_similarities

Cpredicted_rating

Dnp.mean(neighbor_ratings)

Attempts:

3 left

💡 Hint

Common Mistakes

Multiplying similarities by themselves instead of ratings.

Using mean instead of weighted average.

To predict the rating, multiply each neighbor's rating by their similarity and sum these, then divide by sum of similarities.

4fill in blank

hard

Fill both blanks to create a dictionary of item recommendations for a user based on predicted ratings above 3.

ML Python

predicted_ratings = {'item1': 4.5, 'item2': 2.0, 'item3': 3.5, 'item4': 1.0}
recommendations = {item: rating for item, rating in predicted_ratings.items() if rating [1] [2]

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

A>

B3

C<

D4

Attempts:

3 left

💡 Hint

Common Mistakes

Using '<' instead of '>' causing wrong filtering.

Using wrong threshold number.

We want to keep items with ratings greater than 3, so use '>' and 3 in the condition.

5fill in blank

hard

Fill all three blanks to compute user similarity matrix and get top 2 similar users for user 0.

ML Python

from sklearn.metrics.pairwise import [1]
import numpy as np
ratings = np.array([[5, 3, 0], [4, 0, 0], [1, 1, 0], [5, 0, 4]])
sim_matrix = [1](ratings)
top_users = sim_matrix[0].argsort()[::-1][[3]:[2]]

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

Acosine_similarity

B1

C3

D0

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong similarity function.

Incorrect slicing indices for top users.

Use cosine_similarity to compute similarity matrix. argsort()[::-1] sorts descending. To get top 2 excluding self (index 0), slice from 1 to 3.