Back to Subject

Unit 6: Recommender Systems - Practice Quiz

INT423 — Machine Learning-Ii 50 Questions
0 Correct 0 Wrong 50 Left
0/50

1 What is the primary objective of a Recommender System?

A. To predict user preferences for items they have not yet interacted with
B. To reduce the dimensionality of a dataset
C. To cluster similar users without predicting ratings
D. To classify images

2 Which of the following best describes Content-Based Filtering?

A. It uses matrix factorization to find latent features.
B. It recommends items based on the preferences of similar users.
C. It relies solely on demographic data.
D. It recommends items similar to those a user liked in the past based on item features.

3 What is the 'Cold Start' problem in recommender systems?

A. The time it takes to initialize the recommendation engine.
B. The issue where popular items are recommended too often.
C. The system overheats due to large data processing.
D. The difficulty in recommending items to a new user or recommending a new item due to lack of history.

4 Which of the following represents 'Explicit Feedback'?

A. A user watching a video for 10 minutes.
B. A user adding an item to a cart but not buying it.
C. A user giving a movie a 5-star rating.
D. A user clicking on a product link.

5 Which of the following represents 'Implicit Feedback'?

A. Writing a text review.
B. Filling out a preference survey.
C. Clicking on an advertisement.
D. Rating a song 4 out of 5.

6 In Collaborative Filtering, what is the core underlying assumption?

A. The features of the items are more important than user interactions.
B. Users who agreed in the past will tend to agree in the future.
C. Items with similar descriptions are always rated similarly.
D. Users' preferences change randomly over time.

7 What is the purpose of Mean Normalization in collaborative filtering?

A. To remove the effect of item features.
B. To handle users who have not rated any items by treating missing ratings as the average.
C. To convert binary labels into continuous variables.
D. To increase the range of ratings.

8 In a binary label system (favs, likes, clicks), how is the target variable usually represented?

A. As a discrete set {1, 2, 3, 4, 5}.
B. As a vector of text keywords.
C. As a continuous value between 0 and 1.
D. As 1 for interaction (positive) and 0 for no interaction.

9 Which similarity measure is commonly used in Content-Based Filtering to compare document vectors?

A. Hamming Distance
B. Euclidean Distance
C. Manhattan Distance
D. Cosine Similarity

10 What is a major disadvantage of Content-Based Filtering?

A. It cannot handle binary data.
B. It tends to overspecialize and lacks serendipity (surprising recommendations).
C. It suffers from the cold start problem for new users.
D. It requires a large number of users to find similarities.

11 User-based Collaborative Filtering involves:

A. Finding users who live in the same demographic area.
B. Finding users similar to the target user and recommending what they liked.
C. Filtering content based on keywords provided by the user.
D. Finding items similar to the item the user is viewing.

12 Item-based Collaborative Filtering involves:

A. Using a decision tree to classify items.
B. Clustering users into groups.
C. Calculating the similarity between items based on user co-ratings.
D. Analyzing item descriptions to find keywords.

13 What is 'Matrix Factorization' in the context of recommender systems?

A. A way to normalize the mean of the matrix.
B. A method to sort the matrix by top-rated items.
C. A method to multiply two matrices to get the final ratings.
D. A technique to decompose the user-item interaction matrix into lower-dimensional latent factor matrices.

14 When using binary labels like 'clicks', which issue is most prominent compared to explicit ratings?

A. Lack of user identification
B. Ambiguity of negative feedback (missing data vs. dislike)
C. Data Scarcity
D. High computational cost

15 Which of the following is a key advantage of Hybrid Recommender Systems?

A. They only require implicit feedback.
B. They can overcome limitations of individual approaches like the cold start problem.
C. They eliminate the need for data collection.
D. They are computationally cheaper than simple algorithms.

16 In a Recommender System, 'Serendipity' refers to:

A. The accuracy of the prediction.
B. The ability to recommend items that are relevant but surprising to the user.
C. The consistency of recommendations over time.
D. The speed of the recommendation engine.

17 If a user has rated Item A (5 stars) and Item B (5 stars), and a second user rated Item A (5 stars), Item-Based CF would likely predict:

A. The second user is a bot.
B. The second user will give Item B a high rating.
C. The second user will dislike Item B.
D. No prediction is possible.

18 Which technique is best suited for a system where users rarely rate items but generate many search queries?

A. User-based Collaborative Filtering
B. Matrix Factorization on explicit ratings
C. Content-Based Filtering
D. Demographic Filtering

19 What is the 'Long Tail' phenomenon in recommender systems?

A. The system requires long user IDs.
B. Recommendations are presented in a long list.
C. A small number of popular items generate most interactions, while many niche items have few interactions.
D. The algorithm takes a long time to converge.

20 How does Mean Normalization help with bias?

A. It removes users who always give 1-star ratings.
B. It ignores the item bias.
C. It adjusts for users who are consistently harsh or generous in their ratings.
D. It converts all ratings to positive integers.

21 Which algorithm is commonly used to train Matrix Factorization models?

A. Alternating Least Squares (ALS)
B. Apriori Algorithm
C. K-Means Clustering
D. Decision Trees

22 A Weighted Hybrid Recommender System works by:

A. Using content filtering only when collaborative filtering fails.
B. Selecting one algorithm randomly.
C. Combining the scores of different recommendation techniques with specific weights.
D. Running algorithms in a sequence where one refines the other.

23 What is a 'Switching' Hybrid System?

A. It switches between positive and negative ratings.
B. It chooses a recommendation technique based on the current situation (e.g., data availability).
C. It swaps the item ID with the user ID.
D. It switches the user interface based on preferences.

24 In the context of binary labels, what is 'Confidence' often associated with?

A. The percentage of items rated.
B. The probability that the user is a human.
C. The strength of the interaction (e.g., frequency of clicks or duration of view).
D. The confidence interval of the error.

25 What is the primary input for a Content-Based Filtering algorithm?

A. A User-Item Rating Matrix.
B. Social Network Graphs.
C. Demographic data of all users.
D. Item Profiles (Features) and User Profiles.

26 Collaborative Filtering generally outperforms Content-Based Filtering in which scenario?

A. When recommending to a brand new user.
B. When items have rich, structured metadata.
C. When identifying cross-genre or complex patterns that are hard to feature-engineer.
D. When there are no user ratings available.

27 Which metric is commonly used to evaluate a Recommender System utilizing explicit ratings?

A. Jaccard Index
B. F1-Score
C. Accuracy
D. Root Mean Squared Error (RMSE)

28 What is the 'Grey Sheep' problem?

A. Items that are black and white images.
B. Users who only rate popular items.
C. The problem of duplicate accounts.
D. Users whose opinions do not consistently agree or disagree with any group of people.

29 In a user-item matrix used for CF, what does 'Sparsity' refer to?

A. The ratings are all low numbers.
B. Most entries in the matrix are empty (unknown ratings).
C. The matrix is small in size.
D. The matrix has low rank.

30 Which feature engineering technique is essential for Content-Based filtering of text documents?

A. Fourier Transform
B. Pixel normalization
C. Min-Max Scaling of IDs
D. TF-IDF (Term Frequency-Inverse Document Frequency)

31 Latent factors in Matrix Factorization usually represent:

A. The user's age and location.
B. The timestamp of the rating.
C. Hidden characteristics inferred from data patterns.
D. Explicit categories like 'Action' or 'Comedy'.

32 If a dataset consists only of 'Purchase' vs 'Non-Purchase', which type of filtering is applied?

A. Sentiment Analysis
B. Explicit Rating CF
C. Regression Analysis
D. Implicit Feedback CF

33 Why is Mean Normalization important when using Cosine Similarity for centered data (Pearson Correlation)?

A. It speeds up the computation.
B. It removes the user ID from the calculation.
C. It transforms the cosine similarity into Pearson Correlation Coefficient.
D. It ensures all vectors have unit length.

34 Which of the following is a limitation of Collaborative Filtering?

A. It yields recommendations that are too obvious.
B. It cannot recommend items if no one else has rated them (New Item problem).
C. It requires domain knowledge to engineer features.
D. It is strictly rule-based.

35 In a Cascade Hybrid System:

A. One recommender refines the recommendations given by another.
B. All recommenders run in parallel.
C. The system cascades into a random selection.
D. The weights of recommenders change dynamically.

36 Which strategy helps in solving the Cold Start problem for a new user?

A. Asking the user to select preferred genres during onboarding.
B. Waiting for the user to rate 100 items.
C. Using Item-Based Collaborative Filtering.
D. Applying Matrix Factorization immediately.

37 What is the primary advantage of Model-Based CF over Memory-Based CF?

A. It gives exact results based on neighbors.
B. It does not require training.
C. It is easier to implement.
D. It handles sparsity better and offers faster prediction times.

38 What does the 'Banana' problem refer to in Recommender Systems?

A. A coding error in Python.
B. Users buying bananas only once.
C. The shape of the loss function.
D. Recommending items like bananas which are bought frequently but don't indicate distinct taste.

39 Which loss function is minimized in standard Matrix Factorization for explicit ratings?

A. Cross-Entropy Loss
B. Squared Error (between actual and predicted rating) + Regularization
C. Log-Likelihood
D. Hinge Loss

40 How does a 'Demographic-based' recommender work?

A. It uses the text content of reviews.
B. It uses the age, gender, and location of users to find similar groups.
C. It uses only purchase history.
D. It uses satellite imagery.

41 Which of the following is an example of a use case for Association Rule Learning in recommendations?

A. Determining if an email is spam.
B. Market Basket Analysis (e.g., 'Frequently Bought Together').
C. Face recognition.
D. Predicting the rating of a movie.

42 Binary cross-entropy is a suitable loss function when:

A. Clustering users.
B. Predicting a star rating from 1 to 5.
C. Predicting a probability of interaction (click/no-click).
D. Predicting the price of a house.

43 What is the scalability challenge in User-Based Collaborative Filtering (Memory-Based)?

A. It only works with binary data.
B. Computing similarity between millions of users in real-time is computationally expensive.
C. It requires too much hard drive space.
D. It cannot handle text data.

44 In the context of Mean Normalization, if a user has not rated any movies, what prediction does the algorithm default to?

A. The average rating of the specific movie by other users.
B. A random number.
C. Zero.
D. The maximum possible rating.

45 What is 'Collaborative Filtering' distinct from?

A. Using machine learning.
B. Analyzing the internal attributes or content of the item.
C. Using matrices.
D. Predicting future behavior.

46 Regularization is added to the cost function in Matrix Factorization to:

A. Ensure all ratings are positive.
B. Make the code run faster.
C. Increase the number of latent features.
D. Prevent overfitting by penalizing large values in the feature matrices.

47 Which of these is a binary label?

A. View Duration: 120 seconds
B. Rating: 4.5 stars
C. Review Sentiment: Positive (0.8 score)
D. Favorite: Yes

48 Comparison: Which method requires domain knowledge for feature extraction?

A. Content-Based Filtering
B. User-User KNN
C. Collaborative Filtering
D. Matrix Factorization

49 Precision@k is a metric used to evaluate:

A. The time taken to generate k recommendations.
B. The number of users who rated k items.
C. The proportion of recommended items in the top-k set that are relevant.
D. The exact numerical rating accuracy.

50 In a Hybrid system, 'Feature Augmentation' refers to:

A. Using the output of one recommender as a feature input for another.
B. Increasing the font size of recommendations.
C. Adding more RAM to the server.
D. Adding random noise to the data.
Previous MCQ
Unit 5