1What is the primary objective of a Recommender System?
A.To cluster similar users without predicting ratings
B.To predict user preferences for items they have not yet interacted with
C.To classify images
D.To reduce the dimensionality of a dataset
Correct Answer: To predict user preferences for items they have not yet interacted with
Explanation:
The main goal of a recommender system is to estimate the preference or rating a user would give to an item they have not yet seen.
Incorrect! Try again.
2Which of the following best describes Content-Based Filtering?
A.It recommends items based on the preferences of similar users.
B.It recommends items similar to those a user liked in the past based on item features.
C.It relies solely on demographic data.
D.It uses matrix factorization to find latent features.
Correct Answer: It recommends items similar to those a user liked in the past based on item features.
Explanation:
Content-based filtering uses item attributes (features) and a user's profile of preferred features to make recommendations.
Incorrect! Try again.
3What is the 'Cold Start' problem in recommender systems?
A.The issue where popular items are recommended too often.
B.The time it takes to initialize the recommendation engine.
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.
Correct Answer: The difficulty in recommending items to a new user or recommending a new item due to lack of history.
Explanation:
The cold start problem occurs when the system does not have enough data (interactions) for a new user or a new item to make accurate predictions.
Incorrect! Try again.
4Which of the following represents 'Explicit Feedback'?
A.A user watching a video for 10 minutes.
B.A user clicking on a product link.
C.A user adding an item to a cart but not buying it.
D.A user giving a movie a 5-star rating.
Correct Answer: A user giving a movie a 5-star rating.
Explanation:
Explicit feedback is when a user directly specifies their preference, such as a star rating or a like/dislike button.
Incorrect! Try again.
5Which of the following represents 'Implicit Feedback'?
A.Clicking on an advertisement.
B.Writing a text review.
C.Filling out a preference survey.
D.Rating a song 4 out of 5.
Correct Answer: Clicking on an advertisement.
Explanation:
Implicit feedback is inferred from user behavior, such as clicks, purchase history, or watch time, rather than direct ratings.
Incorrect! Try again.
6In Collaborative Filtering, what is the core underlying assumption?
A.Users who agreed in the past will tend to agree in the future.
B.Users' preferences change randomly over time.
C.The features of the items are more important than user interactions.
D.Items with similar descriptions are always rated similarly.
Correct Answer: Users who agreed in the past will tend to agree in the future.
Explanation:
Collaborative filtering assumes that if users had similar tastes in the past (correlated ratings), they will have similar tastes in the future.
Incorrect! Try again.
7What is the purpose of Mean Normalization in collaborative filtering?
A.To convert binary labels into continuous variables.
B.To handle users who have not rated any items by treating missing ratings as the average.
C.To remove the effect of item features.
D.To increase the range of ratings.
Correct Answer: To handle users who have not rated any items by treating missing ratings as the average.
Explanation:
Mean normalization centers ratings around zero for each user or item, ensuring that new users with no history receive valid predictions (usually the item's average rating) rather than a zero score.
Incorrect! Try again.
8In a binary label system (favs, likes, clicks), how is the target variable usually represented?
A.As a continuous value between 0 and 1.
B.As 1 for interaction (positive) and 0 for no interaction.
C.As a vector of text keywords.
D.As a discrete set {1, 2, 3, 4, 5}.
Correct Answer: As 1 for interaction (positive) and 0 for no interaction.
Explanation:
Binary labels distinguish between an event happening (1, e.g., a like or click) and not happening (0).
Incorrect! Try again.
9Which similarity measure is commonly used in Content-Based Filtering to compare document vectors?
A.Manhattan Distance
B.Hamming Distance
C.Euclidean Distance
D.Cosine Similarity
Correct Answer: Cosine Similarity
Explanation:
Cosine similarity is widely used to measure the angle between two feature vectors, effectively ignoring magnitude and focusing on the orientation (similarity of content).
Incorrect! Try again.
10What is a major disadvantage of Content-Based Filtering?
A.It tends to overspecialize and lacks serendipity (surprising recommendations).
B.It cannot handle binary data.
C.It suffers from the cold start problem for new users.
D.It requires a large number of users to find similarities.
Correct Answer: It tends to overspecialize and lacks serendipity (surprising recommendations).
Explanation:
Content-based systems only recommend items similar to what the user has already seen, preventing the discovery of different types of content (lack of serendipity).
Incorrect! Try again.
11User-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.
Correct Answer: Finding users similar to the target user and recommending what they liked.
Explanation:
User-based CF identifies users with similar rating histories to the target user and suggests items that those similar users liked.
Incorrect! Try again.
12Item-based Collaborative Filtering involves:
A.Using a decision tree to classify items.
B.Clustering users into groups.
C.Analyzing item descriptions to find keywords.
D.Calculating the similarity between items based on user co-ratings.
Correct Answer: Calculating the similarity between items based on user co-ratings.
Explanation:
Item-based CF recommends items that are similar to items the user has liked in the past, where similarity is calculated based on how other users rated those items.
Incorrect! Try again.
13What 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 technique to decompose the user-item interaction matrix into lower-dimensional latent factor matrices.
D.A method to multiply two matrices to get the final ratings.
Correct Answer: A technique to decompose the user-item interaction matrix into lower-dimensional latent factor matrices.
Explanation:
Matrix Factorization (like SVD) breaks the sparse user-item matrix into two smaller matrices (user factors and item factors) that approximate the original interactions.
Incorrect! Try again.
14When using binary labels like 'clicks', which issue is most prominent compared to explicit ratings?
A.High computational cost
B.Ambiguity of negative feedback (missing data vs. dislike)
C.Lack of user identification
D.Data Scarcity
Correct Answer: Ambiguity of negative feedback (missing data vs. dislike)
Explanation:
With binary implicit feedback (clicks), a '0' could mean the user dislikes the item or simply hasn't seen it yet; explicit dislike is rarely captured.
Incorrect! Try again.
15Which of the following is a key advantage of Hybrid Recommender Systems?
A.They are computationally cheaper than simple algorithms.
B.They can overcome limitations of individual approaches like the cold start problem.
C.They only require implicit feedback.
D.They eliminate the need for data collection.
Correct Answer: They can overcome limitations of individual approaches like the cold start problem.
Explanation:
Hybrid systems combine techniques (e.g., Content-based and CF) to mitigate the weaknesses of one method with the strengths of another.
Incorrect! Try again.
16In a Recommender System, 'Serendipity' refers to:
A.The accuracy of the prediction.
B.The consistency of recommendations over time.
C.The ability to recommend items that are relevant but surprising to the user.
D.The speed of the recommendation engine.
Correct Answer: The ability to recommend items that are relevant but surprising to the user.
Explanation:
Serendipity is the quality of finding valuable or pleasant things that were not looked for; in recommendations, it means suggesting items the user might not have discovered on their own.
Incorrect! Try again.
17If 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 will give Item B a high rating.
B.No prediction is possible.
C.The second user is a bot.
D.The second user will dislike Item B.
Correct Answer: The second user will give Item B a high rating.
Explanation:
Since the items share similar rating patterns from the first user, Item-Based CF assumes the second user will also rate Item B highly.
Incorrect! Try again.
18Which technique is best suited for a system where users rarely rate items but generate many search queries?
A.User-based Collaborative Filtering
B.Content-Based Filtering
C.Matrix Factorization on explicit ratings
D.Demographic Filtering
Correct Answer: Content-Based Filtering
Explanation:
Search queries provide feature data (keywords) that align well with Content-Based Filtering, which relies on item/content attributes rather than rating history.
Incorrect! Try again.
19What is the 'Long Tail' phenomenon in recommender systems?
A.The system requires long user IDs.
B.The algorithm takes a long time to converge.
C.A small number of popular items generate most interactions, while many niche items have few interactions.
D.Recommendations are presented in a long list.
Correct Answer: A small number of popular items generate most interactions, while many niche items have few interactions.
Explanation:
The Long Tail refers to the large volume of niche items that are individually unpopular but collectively significant. Recommender systems help users discover these items.
Incorrect! Try again.
20How does Mean Normalization help with bias?
A.It removes users who always give 1-star ratings.
B.It converts all ratings to positive integers.
C.It adjusts for users who are consistently harsh or generous in their ratings.
D.It ignores the item bias.
Correct Answer: It adjusts for users who are consistently harsh or generous in their ratings.
Explanation:
By subtracting the user's average rating from their specific ratings, the system analyzes deviations from their personal norm, neutralizing the effect of being a 'harsh' or 'generous' rater.
Incorrect! Try again.
21Which algorithm is commonly used to train Matrix Factorization models?
A.K-Means Clustering
B.Alternating Least Squares (ALS)
C.Decision Trees
D.Apriori Algorithm
Correct Answer: Alternating Least Squares (ALS)
Explanation:
ALS is a popular optimization algorithm for Matrix Factorization, especially when dealing with large, sparse datasets and implicit feedback.
Incorrect! Try again.
22A Weighted Hybrid Recommender System works by:
A.Combining the scores of different recommendation techniques with specific weights.
B.Using content filtering only when collaborative filtering fails.
C.Running algorithms in a sequence where one refines the other.
D.Selecting one algorithm randomly.
Correct Answer: Combining the scores of different recommendation techniques with specific weights.
Explanation:
A weighted hybrid computes the final recommendation score as a weighted sum (or linear combination) of scores from multiple individual recommenders.
Incorrect! Try again.
23What is a 'Switching' Hybrid System?
A.It switches the user interface based on preferences.
B.It chooses a recommendation technique based on the current situation (e.g., data availability).
C.It switches between positive and negative ratings.
D.It swaps the item ID with the user ID.
Correct Answer: It chooses a recommendation technique based on the current situation (e.g., data availability).
Explanation:
Switching hybrids select a single algorithm to use for a specific instance based on criteria like confidence level or data availability (e.g., use Content-based for new users, CF for existing ones).
Incorrect! Try again.
24In the context of binary labels, what is 'Confidence' often associated with?
A.The probability that the user is a human.
B.The percentage of items rated.
C.The confidence interval of the error.
D.The strength of the interaction (e.g., frequency of clicks or duration of view).
Correct Answer: The strength of the interaction (e.g., frequency of clicks or duration of view).
Explanation:
In implicit feedback (binary) systems, repeated interactions (like listening to a song 10 times) increase the 'confidence' that the binary '1' represents a true positive preference.
Incorrect! Try again.
25What is the primary input for a Content-Based Filtering algorithm?
A.Demographic data of all users.
B.A User-Item Rating Matrix.
C.Social Network Graphs.
D.Item Profiles (Features) and User Profiles.
Correct Answer: Item Profiles (Features) and User Profiles.
Explanation:
Content-based filtering relies on the attributes of the items (Item Profiles) and the attributes preferred by the user (User Profiles).
Incorrect! Try again.
26Collaborative Filtering generally outperforms Content-Based Filtering in which scenario?
A.When identifying cross-genre or complex patterns that are hard to feature-engineer.
B.When there are no user ratings available.
C.When items have rich, structured metadata.
D.When recommending to a brand new user.
Correct Answer: When identifying cross-genre or complex patterns that are hard to feature-engineer.
Explanation:
CF can capture complex behavioral patterns (e.g., people who buy beer also buy diapers) that feature-based content analysis might miss.
Incorrect! Try again.
27Which metric is commonly used to evaluate a Recommender System utilizing explicit ratings?
A.Accuracy
B.Root Mean Squared Error (RMSE)
C.F1-Score
D.Jaccard Index
Correct Answer: Root Mean Squared Error (RMSE)
Explanation:
RMSE measures the difference between predicted ratings and actual ratings, making it the standard metric for regression-type recommendation tasks.
Incorrect! Try again.
28What is the 'Grey Sheep' problem?
A.Users who only rate popular items.
B.Items that are black and white images.
C.The problem of duplicate accounts.
D.Users whose opinions do not consistently agree or disagree with any group of people.
Correct Answer: Users whose opinions do not consistently agree or disagree with any group of people.
Explanation:
The Grey Sheep problem refers to users with unique or inconsistent tastes, making it difficult for Collaborative Filtering to find similar users to generate recommendations.
Incorrect! Try again.
29In a user-item matrix used for CF, what does 'Sparsity' refer to?
A.The matrix is small in size.
B.The matrix has low rank.
C.Most entries in the matrix are empty (unknown ratings).
D.The ratings are all low numbers.
Correct Answer: Most entries in the matrix are empty (unknown ratings).
Explanation:
Sparsity means that out of all possible user-item combinations, very few actual interactions have occurred, resulting in a matrix filled mostly with empty values.
Incorrect! Try again.
30Which feature engineering technique is essential for Content-Based filtering of text documents?
TF-IDF is used to evaluate the importance of a word in a document relative to a collection of documents, creating a feature vector for text-based items.
Incorrect! Try again.
31Latent factors in Matrix Factorization usually represent:
A.The user's age and location.
B.Hidden characteristics inferred from data patterns.
C.The timestamp of the rating.
D.Explicit categories like 'Action' or 'Comedy'.
Correct Answer: Hidden characteristics inferred from data patterns.
Explanation:
Latent factors are abstract dimensions found by the algorithm that explain variance in ratings; they may correlate with genres or styles but are not explicitly labeled.
Incorrect! Try again.
32If a dataset consists only of 'Purchase' vs 'Non-Purchase', which type of filtering is applied?
A.Explicit Rating CF
B.Sentiment Analysis
C.Implicit Feedback CF
D.Regression Analysis
Correct Answer: Implicit Feedback CF
Explanation:
Purchase history is a form of implicit feedback (binary behavior), as opposed to explicit star ratings.
Incorrect! Try again.
33Why is Mean Normalization important when using Cosine Similarity for centered data (Pearson Correlation)?
A.It ensures all vectors have unit length.
B.It transforms the cosine similarity into Pearson Correlation Coefficient.
C.It removes the user ID from the calculation.
D.It speeds up the computation.
Correct Answer: It transforms the cosine similarity into Pearson Correlation Coefficient.
Explanation:
Cosine similarity on mean-centered data is mathematically equivalent to the Pearson Correlation Coefficient.
Incorrect! Try again.
34Which of the following is a limitation of Collaborative Filtering?
A.It cannot recommend items if no one else has rated them (New Item problem).
B.It is strictly rule-based.
C.It yields recommendations that are too obvious.
D.It requires domain knowledge to engineer features.
Correct Answer: It cannot recommend items if no one else has rated them (New Item problem).
Explanation:
Since CF relies on past interactions, a new item with zero interactions cannot be recommended until it is rated by someone.
Incorrect! Try again.
35In a Cascade Hybrid System:
A.All recommenders run in parallel.
B.One recommender refines the recommendations given by another.
C.The weights of recommenders change dynamically.
D.The system cascades into a random selection.
Correct Answer: One recommender refines the recommendations given by another.
Explanation:
In a cascade hybrid, a primary technique produces a candidate set, and a secondary technique refines or re-ranks this set.
Incorrect! Try again.
36Which strategy helps in solving the Cold Start problem for a new user?
A.Asking the user to select preferred genres during onboarding.
B.Applying Matrix Factorization immediately.
C.Waiting for the user to rate 100 items.
D.Using Item-Based Collaborative Filtering.
Correct Answer: Asking the user to select preferred genres during onboarding.
Explanation:
Gathering initial preferences allows the system to use demographic or content-based strategies before interaction history is available.
Incorrect! Try again.
37What is the primary advantage of Model-Based CF over Memory-Based CF?
A.It does not require training.
B.It is easier to implement.
C.It handles sparsity better and offers faster prediction times.
D.It gives exact results based on neighbors.
Correct Answer: It handles sparsity better and offers faster prediction times.
Explanation:
Model-based CF (like SVD) compresses data into a model, which handles sparse data better and predicts faster than scanning the entire database (Memory-based) at runtime.
Incorrect! Try again.
38What does the 'Banana' problem refer to in Recommender Systems?
A.Recommending items like bananas which are bought frequently but don't indicate distinct taste.
B.Users buying bananas only once.
C.A coding error in Python.
D.The shape of the loss function.
Correct Answer: Recommending items like bananas which are bought frequently but don't indicate distinct taste.
Explanation:
This refers to the issue of grocery recommendations where frequent purchases (staples like bananas) might crowd out discovery of interesting, taste-based items.
Incorrect! Try again.
39Which loss function is minimized in standard Matrix Factorization for explicit ratings?
A.Cross-Entropy Loss
B.Log-Likelihood
C.Squared Error (between actual and predicted rating) + Regularization
D.Hinge Loss
Correct Answer: Squared Error (between actual and predicted rating) + Regularization
Explanation:
Matrix factorization aims to minimize the squared difference between the dot product of latent factors and the actual rating, usually with a regularization term to prevent overfitting.
Incorrect! Try again.
40How does a 'Demographic-based' recommender work?
A.It uses only purchase history.
B.It uses the age, gender, and location of users to find similar groups.
C.It uses satellite imagery.
D.It uses the text content of reviews.
Correct Answer: It uses the age, gender, and location of users to find similar groups.
Explanation:
Demographic filtering assumes that users with similar demographic attributes (age, location, etc.) will have similar preferences.
Incorrect! Try again.
41Which of the following is an example of a use case for Association Rule Learning in recommendations?
Association rules (like Apriori) identify items that frequently co-occur in transactions, typically used for 'bundle' recommendations.
Incorrect! Try again.
42Binary cross-entropy is a suitable loss function when:
A.Predicting a probability of interaction (click/no-click).
B.Clustering users.
C.Predicting the price of a house.
D.Predicting a star rating from 1 to 5.
Correct Answer: Predicting a probability of interaction (click/no-click).
Explanation:
Binary cross-entropy is the standard loss function for binary classification problems, such as predicting whether a user will click (1) or not (0).
Incorrect! Try again.
43What is the scalability challenge in User-Based Collaborative Filtering (Memory-Based)?
A.It cannot handle text data.
B.Computing similarity between millions of users in real-time is computationally expensive.
C.It only works with binary data.
D.It requires too much hard drive space.
Correct Answer: Computing similarity between millions of users in real-time is computationally expensive.
Explanation:
Memory-based approaches require calculating distances between the target user and all other users, which becomes slow (O(N^2) or O(NM)) as user count grows.
Incorrect! Try again.
44In the context of Mean Normalization, if a user has not rated any movies, what prediction does the algorithm default to?
A.Zero.
B.The average rating of the specific movie by other users.
C.The maximum possible rating.
D.A random number.
Correct Answer: The average rating of the specific movie by other users.
Explanation:
With mean normalization, a user's bias is 0 if undefined. The prediction becomes (User Bias + Movie Average). If User Bias is 0, the prediction is just the Movie Average.
Incorrect! Try again.
45What is 'Collaborative Filtering' distinct from?
A.Using matrices.
B.Analyzing the internal attributes or content of the item.
C.Using machine learning.
D.Predicting future behavior.
Correct Answer: Analyzing the internal attributes or content of the item.
Explanation:
This is the defining difference: CF ignores what the item is (content) and focuses only on how users interact with it.
Incorrect! Try again.
46Regularization is added to the cost function in Matrix Factorization to:
A.Increase the number of latent features.
B.Prevent overfitting by penalizing large values in the feature matrices.
C.Ensure all ratings are positive.
D.Make the code run faster.
Correct Answer: Prevent overfitting by penalizing large values in the feature matrices.
Explanation:
Regularization terms (like Lambda * magnitude) prevent the learned parameters from becoming too large and fitting the noise in the training data.
Incorrect! Try again.
47Which of these is a binary label?
A.Review Sentiment: Positive (0.8 score)
B.View Duration: 120 seconds
C.Rating: 4.5 stars
D.Favorite: Yes
Correct Answer: Favorite: Yes
Explanation:
A 'Favorite' action is a binary state: an item is either favorited (1) or it is not (0).
Incorrect! Try again.
48Comparison: Which method requires domain knowledge for feature extraction?
A.User-User KNN
B.Matrix Factorization
C.Content-Based Filtering
D.Collaborative Filtering
Correct Answer: Content-Based Filtering
Explanation:
Content-based filtering requires understanding the items to extract relevant features (e.g., genre, director, bpm, keywords), requiring domain knowledge.
Incorrect! Try again.
49Precision@k is a metric used to evaluate:
A.The exact numerical rating accuracy.
B.The number of users who rated k items.
C.The time taken to generate k recommendations.
D.The proportion of recommended items in the top-k set that are relevant.
Correct Answer: The proportion of recommended items in the top-k set that are relevant.
Explanation:
Precision@k measures how many of the top 'k' recommendations presented to the user were actually relevant (interacted with).
Incorrect! Try again.
50In a Hybrid system, 'Feature Augmentation' refers to:
A.Increasing the font size of recommendations.
B.Adding more RAM to the server.
C.Adding random noise to the data.
D.Using the output of one recommender as a feature input for another.
Correct Answer: Using the output of one recommender as a feature input for another.
Explanation:
Feature augmentation involves generating a rating or classification using one algorithm and feeding that result as a feature into the next algorithm.