NEW
Font size
WorksheetsRecommendation Systems Quiz
Total questions: 30
Worksheet time: 15mins
What is the main goal of a recommendation system?
Store user data
Recommend items to users
Delete user preferences
Translate languages
Which of the following is a content-based recommendation example?
Users who liked this also liked that
Movies similar to Inception
Top trending movies
Most purchased products
Which library in Python is commonly used to handle datasets?
NumPy
SciPy
pandas
Matplotlib
Which technique is used to convert text into numbers in content-based filtering?
Word2Vec
TF-IDF
CNN
Naive Bayes
TF-IDF stands for:
Text Frequency–Inverted Data Factor
Term Frequency–Inverse Document Frequency
Text Format–Index Data File
Term Format–Inverse Data Feature
What type of filtering uses user-item interactions?
Content-based
Collaborative
Rule-based
Matrix-based
What function from sklearn is used to find similarity between items?
cosine_curve()
cosine_similarity()
similarity_matrix()
distance_score()
In a content-based system, the comparison is made between:
Users
Items
Users and Items
Products and Prices
The output of cosine similarity ranges between:
-1 and 1
0 and 100
0 and 1
-100 and 100
Which stop words are usually removed by the TF-IDF vectorizer?
Important words
Common words like 'the', 'is', 'and'
Proper nouns
Numeric words
What is the main limitation of a content-based recommender system?
Cold start for new users
It can’t recommend unseen items
It needs large user data
It doesn’t use machine learning
In the TF-IDF formula, what does 'IDF' represent?
Frequency of a word in the same document
Frequency of a word across all documents
Frequency of the least common word
Importance of long documents
In cosine similarity, when two vectors are identical, their cosine value is:
0
0.5
1
-1
What does fit_transform() do in TF-IDF?
Only converts text to lowercase
Learns the vocabulary and transforms text
Normalizes the vectors
Removes duplicates
Which Python library provides TfidfVectorizer?
pandas
sklearn
numpy
tensorflow
What is the purpose of sorting similarity scores in descending order?
To get least similar items
To get most similar items
To shuffle data
To remove duplicates
If the cosine similarity between Movie A and Movie B is 0.95, what does it mean?
Movies are completely different
Movies are very similar
Movies are partially similar
Movies are not comparable
Which method can be used to extend a content-based system for multiple attributes?
Using only text
Combining genre, director, and actors
Ignoring non-text data
Using single feature
Which one is a disadvantage of TF-IDF?
It can’t understand word meaning or context
It’s very slow
It doesn’t work with English
It removes punctuation automatically
Which data structure stores all similarity scores in this system?
List
Matrix
Dictionary
Tuple
What happens if two movie descriptions have no overlapping words?
Cosine similarity = 1
Cosine similarity = 0
Cosine similarity = -1
Cosine similarity = 0.5
What is the dimension of the cosine similarity matrix if we have 5 movies?
1x5
5x1
5x5
10x5
In a recommender function, why is the first item skipped ([1:4])?
It’s already the most similar to itself
It has the least similarity
It’s a duplicate record
It has missing data
What would improve semantic understanding beyond TF-IDF?
Random Forest
Word Embeddings (Word2Vec/BERT)
Naive Bayes
PCA
Which metric other than cosine similarity can be used for distance-based comparison?
Euclidean Distance
RMSE
Accuracy
Precision
Why do we use stop words in TF-IDF Vectorizer?
To keep unnecessary words
To ignore unimportant words
To increase vocabulary
To reduce document count
Which part of the content-based recommendation can lead to over-specialization?
Relying only on content similarity
Using cosine similarity
Using hybrid models
Using user feedback
If a movie has a description of only one word repeated many times, TF-IDF will:
Assign very high score to that word
Reduce its weight due to frequency
Ignore the word completely
Normalize to zero
Which Python function in sklearn helps remove frequently used words?
stop_words='english'
remove_words=True
clear_common=True
stop_common_words()
In the recommend() function, which line selects top 3 similar movies?
sim_scores = list(enumerate(cosine_sim[idx]))
sim_scores = sim_scores[1:4]
movie_indices = [i[0] for i in sim_scores]
sorted(sim_scores, key=lambda x: x[1])
