WorksheetsMachine Learning Quiz
Total questions: 59
Worksheet time: 30mins
Hey there! Can you guess what we call the application of machine learning methods to analyze customer behavior?
data mining
artificial intelligence
big data computing
internet of things
Hey there! Viaan is working on a cool machine learning project. If Viaan's model uses the target variable in its output, what kind of model is it?
Informative model
Fortune-telling model
Adventure model
All of the above
Hey there! Can you tell me in which type of learning we use labelled training data?
unsupervised learning
supervised learning
reinforcement learning
active learning
Imagine Rahul, Asher, and Sanya embarking on an exciting project in their classroom! They are exploring different feature selection methods. Which method involves starting with an empty feature set?
Forward feature selection
Backward feature selection
Both A and B
None of the above
Hey there! Can you guess if the statement below is true or false?
In PCA the number of input dimensions are equal to principal components
True
False
Avani, Aanya, and Aarav are planning a fun project together. They want to use PCA to reduce the dimensions of their data and visualize it effectively. Is this statement true or false?
true
false
Hey there! Can you help Krish, Divya, and Akhil choose the ultimate machine learning method?
scalable
accuracy
fast
all of the above
Hey there! Can you tell me what makes unlabeled examples stand out in the world of machine learning?
They come with no prior knowledge
They come with no confusing knowledge
They come with some prior knowledge
They come with a lot of confusing knowledge
Hey there! Can you tell me what dimensionality reduction helps with?
stochastics
collinerity
performance
entropy
Hey there! Can you help Rahul with this exciting challenge? What type of data is used to build a data mining model in a machine learning project?
training data
validation data
test data
hidden data
Avani, Alisha, and Nikita are like the three musketeers of a school. Each one has their unique style and personality, making them stand out from each other. Just like how all principal components are orthogonal to each other!
Avani & Alisha
Alisha & Nikita
Nikita & Avani
all of the above
Hey there! Can you help Advait figure out the best way to choose the number of principal components "k"?
Go for the smallest value of k that retains at least 99% of the variance.
Opt for k to be 99% of m (k = 0.99*m, rounded to the nearest integer).
Pick the largest value of k that retains 99% of the variance.
Try using the elbow method.
Avni, Ria, and Aanya are working on a data science project. They need to identify an example of feature extraction. Which of the following actions would be considered as feature extraction?
Constructing bag of words from an email
Applying PCA to reduce high dimensional data
Removing stop words
Performing forward selection
Hey there! Let's have some fun with a question:
What is the definition of Prediction?
It's like magic - the result of applying a specific theory or rule in a specific case
Imagine it as a cool tool in statistics that helps find projections in multidimensional data
Think of it as a secret code entered in a database by an expert
Picture it as something that stands on its own, independent of data
Hey there! Riyaan, Anika, and Neha are exploring weather data to predict the next storm. What type of learning is this an example of?
supervised learning
reinforcement learning
unsupervised learning
dimensionality reduction
Hey there! Let's play a game of knowledge. Can you tell me when PCA works best? 1. When there is a linear structure in the data 2. When the data lies on a curved surface and not on a flat surface 3. When variables are scaled in the same unit
1 and 2
2 and 3
1 and 3
1,2 and 3
Siya, a brilliant student, has a magical Grade variable F1 that can transform into A, B, C, or D. Can you guess which statement is correct?
F1 is like a secret code, a nominal variable!
F1 is like a ranking system, an ordinal variable!
None of the above, F1 is unique!
F1 is both a ranking system and a secret code!
Hey there! Imagine Tisha, Eshaan, and Arnav are competing in a fun data science challenge. What could be a potential hiccup when trying to implement Leave-One-Out-Cross-Validation (LOOCV) in a real-world scenario?
low variance
high variance
faster runtime compared to k-fold cross validation
slower runtime compared to normal validation
Which model represents Support Vector Machine?
Logical Model
Probabilistic Model
Geometric Model
None of the above
Imagine Ishaan is training a Support Vector Machine model with a super high Gamma value using the RBF kernel. What does this indicate?
the model would consider even far away points from the hyperplane for modeling
the model would consider only the points close to the hyperplane for modeling
the model would not be affected by distance of points from hyperplane for modeling
none of the above
Hey there! Let's talk about a cool trick we use before unleashing the power of the Gaussian kernel in SVM. What's the deal with feature normalization?
It magically scales the features to a standard range
It's like a secret sauce that boosts the model's performance
It's a must-have for tree-based models to shine
None of the above, it's a game-changer!
Hey there! Let's talk about wrapper methods, shall we?
Definitely! They are like the superheroes of hyper-parameter selection methods.
Maybe not always the best choice, but they have their moments of glory.
They are like secret agents, working behind the scenes with 'black box' learning machines.
Let's just say, they are not always the life of the party.
Hey there! Can you guess which superhero can't achieve zero training error on any linearly separable dataset?
Iron Man
Spider-Man
Superman
Batman
Imagine Vanya is on a mission to train a hard-margin linear SVM on a group of n > 100 data points in R2. The hyperplane formed has exactly 2 support vectors. If Aditi sneaks in one more data point and retrain the classifier, what is the maximum possible number of support vectors for the new hyperplane (assuming the total n + 1 points are linearly separable)?
2
3
n
n+1
Let Neha and Divya be the set of magic wands and Rahul and Neha be the enchanted spells for a wizardry problem using different spells respectively. Which of the following are correct?
Neha ⊆ Divya
Neha may not be a subset of Divya
Rahul = Neha
all of the above
Hey there! Let's talk about outliers. Which statement below is true?
Outliers should be part of the training dataset but should not be present in the test data
Outliers should be identified and removed from a dataset
The nature of the problem determines how outliers are used
Outliers should be part of the test dataset but should not be present in the training data
Hey there! Neha, Viaan, and Eesha decided to conduct a fun survey with 105 participants. Among them, 9 were mistakenly identified as positive, 6 as negative, 26 as negative, and 70 as negative. What do you think would be the error rate?
45 percentage
99 percentage
28 percentage
20 percentage
Imagine, Eshaan is exploring a dataset with highly imbalanced class. The majority class is seen 99% of times in the training data. Eshaan's model achieves 99% accuracy after making predictions on test data. What can we conclude in this scenario?
The model is excelling
The model might be overfitting
Accuracy may not be a reliable metric here
None of the above
During a data science competition, Ria is tackling a project that involves training a Support Vector Machine (SVM). The time complexity for training an SVM is at least O(n^2). Based on this information, which dataset sizes are not ideal for SVMs?
large datasets
small datasets
medium sized datasets
size does not matter
Hey there! Can you guess what type of learner Perceptron Classifier is?
Aarav: unsupervised learning algorithm
Tisha: semi-supervised learning algorithm
Aanya: supervised learning algorithm
None of the above
Why did Tara decide to implement PCA in her data science project?
To simplify the data without losing important information
To make the model more sophisticated and accurate
To add a touch of mystery and intrigue to the dataset
None of the above
Hey there! Can you guess which application of SVM is most commonly used?
Image classification
Speech recognition
Text summarization
None of the above
Hey there! Can you guess the top reason why decision trees are awesome in machine learning?
Super easy to interpret and visualize
Exclusively for linear data
Demands feature normalization
None of the above
Hey there! What's the coolest thing about decision trees in machine learning?
They are super easy to interpret and visualize
They are the best for linear data
No need for feature normalization
None of the above, of course!
Why did Aisha decide to implement PCA in her data science project?
To simplify the data representation
To add complexity to the model
To inject some randomness into the dataset
None of the above
Hey there! Can you guess which application of SVM is most common?
Image classification
Speech recognition
Text summarization
None of the above
Why does Aanya love using K-means clustering for customer segmentation in her marketing projects?
Because it helps her group customers based on similar characteristics for targeted marketing campaigns
Because it allows her to analyze customer behavior over time
Because it enables her to predict future customer trends
None of the above
Hey there! Can you help Aanya figure out which of the following is a common application of Random Forest?
Image classification
Speech recognition
Text summarization
None of the above
Hey there! Can you guess the top reason why logistic regression is a superstar in marketing analytics?
It gives you probabilities for outcomes
It's the go-to for linear data
Feature normalization is a must
None of the above
Hey there! Krish is exploring the world of marketing analytics and wants to know - What's the coolest thing about logistic regression?
It gives probabilities for outcomes
It's the go-to for linear data
No need for feature normalization
None of the above
Hey there! Can you tell me the primary purpose of Principal Component Analysis (PCA) in the exciting world of data science?
To simplify the data by reducing its dimensions
To make the model more intricate and interesting
To add some spice by introducing noise in the dataset
None of the above
Hey there! Can you guess which application of SVM is commonly used?
Image classification
Speech recognition
Text summarization
None of the above
Hey there! Can you guess the superpower of logistic regression in the world of machine learning?
Unveils probabilities for outcomes
Exclusively for linear data
Demands feature normalization
None of the above
Hey there! Saisha, Prisha, and Myra are on a mission to understand K-means clustering in machine learning. Can you help them out by selecting the main purpose of using K-means clustering?
To classify data into distinct groups based on similarity
To reduce the dimensionality of the data
To introduce noise in the dataset
None of the above
Hey there! Can you help Aisha figure out which of the following methods won't guarantee zero training error on any linearly separable dataset?
decision tree
15-nearest neighbors
hard-margin svm
perceptron
Hey there! What's the coolest thing about Naive Bayes in the world of machine learning?
Magically handles missing values like a pro
No need for feature scaling, it's a breeze
Works its magic on all types of data, not just linear
Serves up probabilities for outcomes like a fortune teller
Hey there! Avni, Viaan, and Neha are planning a fun activity. Can you guess which one of the following is a common application of Random Forest?
Avni's amazing Image classification skills
Viaan's outstanding Speech recognition abilities
Neha's fantastic Text summarization talent
None of the above
Hey there! Can you tell me what makes K-nearest neighbors so awesome?
Handles missing values like a pro
Needs feature scaling to shine
Super easy to understand and explain
Can dish out probabilities for outcomes
Avni, Asher, and Aisha decided to play a fun game where they conducted a survey with 105 participants. Out of these, 9 were mistakenly identified as positive, 6 were mistakenly identified as negative, 26 were mistakenly identified as negative, and 70 were correctly identified as negative. What would be the error rate?
45 percentage
99 percentage
28 percentage
20 percentage
Hey there! Can you guess the superpower of using logistic regression in machine learning?
Unlocks probabilities for outcomes
Works like a charm for linear data
Demands feature normalization
Sorry, none of the above
Which of the following is a common application of Random Forest?
Image classification
Speech recognition
Text summarization
None of the above
What is the main advantage of using K-nearest neighbors in machine learning?
Simple to implement and understand
Only suitable for linear data
Require feature normalization
None of the above
Which of the following is a common application of Random Forest?
Image classification
Speech recognition
Text summarization
None of the above
What is the main advantage of using ensemble methods in machine learning?
Handles missing values well
Requires feature scaling
Easy to interpret and explain
Provides probabilities for outcomes
Suppose Anika is training a hard-margin linear SVM on n > 100 data points in R2, yielding a hyperplane with exactly 2 support vectors. If Vanya adds one more data point and retrain the classifier, what is the maximum possible number of support vectors for the new hyperplane (assuming the n + 1 points are linearly separable)?
2
3
n
n+1
Why is K-means clustering commonly used in customer segmentation for marketing purposes?
To group customers based on similar characteristics for targeted marketing campaigns
To analyze customer behavior over time
To predict future customer trends
None of the above
Hey there! Can you help Aisha understand which of the following methods is not guaranteed to achieve zero training error on any linearly separable dataset?
decision tree
15-nearest neighbors
hard-margin svm
perceptron
Hey there! Can you explain in which type of learning we utilize labelled training data?
unsupervised learning
supervised learning
reinforcement learning
active learning
Hey there! Can you describe what makes K-nearest neighbors an excellent choice?
Handles missing values like a pro
Needs feature scaling to shine
Super easy to understand and explain
Can dish out probabilities for outcomes
