User-based vs item-based in ML Python - Interactive Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to calculate user similarity using cosine similarity.

ML Python

from sklearn.metrics.pairwise import [1]
user_similarity = [1](user_item_matrix)

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

Apairwise_distances

Bcosine_similarity

Cmanhattan_distances

Deuclidean_distances

Attempts:

3 left

💡 Hint

Common Mistakes

Using distance metrics instead of similarity metrics.

Confusing cosine similarity with Euclidean distance.

The cosine similarity function measures similarity between users based on their item ratings.

2fill in blank

medium

Complete the code to find the top 3 similar items for a given item index.

ML Python

import numpy as np
item_similarities = item_similarity_matrix[item_index]
top_items = np.argsort(item_similarities)[[1]][:3]

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

A-3:

B:3

C::-1

D3:

Attempts:

3 left

💡 Hint

Common Mistakes

Not reversing the sorted indices, resulting in least similar items.

Using incorrect slicing that excludes top items.

Using [::-1] reverses the sorted indices to get the most similar items first.

3fill in blank

hard

Fix the error in the code to compute predicted rating using user-based collaborative filtering.

ML Python

predicted_rating = sum(user_similarity[user_id, [1]] * ratings[[1], item_id]) / sum(abs(user_similarity[user_id, [1]]))

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

Aneighbor_id

Buser_id

Citem_id

Drating_id

Attempts:

3 left

💡 Hint

Common Mistakes

Using user_id instead of neighbor_id causing self-similarity errors.

Using item_id which is unrelated to user similarity indexing.

The neighbor_id represents other users whose similarity and ratings are used to predict the rating.

4fill in blank

hard

Fill both blanks to create an item-based collaborative filtering prediction formula.

ML Python

predicted_rating = sum(item_similarity[item_id, [1]] * ratings[user_id, [1]]) / sum(abs(item_similarity[item_id, [2]]))

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

Aneighbor_item

Buser_id

Ditem_id

Attempts:

3 left

💡 Hint

Common Mistakes

Mixing user_id and item_id in similarity indexing.

Using inconsistent variable names for neighbor items.

Both blanks refer to neighbor items to weight the user's ratings for prediction.

5fill in blank

hard

Fill all three blanks to complete the code that builds a user-item rating matrix from raw data.

ML Python

import pandas as pd
ratings_df = pd.DataFrame(data, columns=[[1], [2], [3]])
rating_matrix = ratings_df.pivot(index=[1], columns=[2], values='rating').fillna(0)

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

A'user_id'

B'item_id'

C'rating'

D'timestamp'

Attempts:

3 left

💡 Hint

Common Mistakes

Using timestamp instead of rating for values.

Swapping user_id and item_id in pivot parameters.

The DataFrame columns must be user_id, item_id, and rating to build the matrix correctly.