NEW
Font size
WorksheetsQuiz 3
Total questions: 20
Worksheet time: 10mins
What does supervised learning use to train a model?
Only input features
Random guesses
Labeled data (input + output)
No data at all
Which one of these is not a performance metric?
Precision
Accuracy
Learning rate
F1 Score
What is the goal of training in machine learning?
To memorize the training data
To minimize the loss or error
To maximize the input size
To increase randomness
What does the law of total probability help you calculate?
Gradient
Probability of combined outcomes
Number of features
Decision boundary
Bayes' Rule helps us update:
Data types
Old models
Probabilities based on new evidence
Regression lines
The dot product of two vectors gives:
A matrix
The angle between them
A scalar
Their average
If a matrix has a non-zero determinant, it means:
It’s empty
It has no inverse
It is invertible
It is orthogonal
What does the rank of a matrix represent?
Its file size
Number of columns
Number of linearly independent rows/columns
How often it’s used
Two vectors are orthogonal if:
They are equal
Their dot product is zero
They have the same direction
They both contain 0
A vector space must always contain:
Only positive vectors
Only unit vectors
The zero vector
At least 10 vectors
Which model uses a tree-like structure to make decisions?
Logistic regression
Decision tree
K-means
Neural network
What does logistic regression output?
A number between 0 and 1
A string
A matrix
A class of clusters
In K-Nearest Neighbors, k represents:
Number of features
Number of classes
Number of closest neighbors to consider
Number of trees
The sigmoid function is used in:
Decision trees
K-means
Logistic regression
Linear regression
What is the main goal of an SVM?
Guess the labels
Use random weights
Find the best margin to separate classes
Increase loss
What is the main idea of clustering?
Label all the data manually
Predict the next number
Group similar data points
Compress the file size
Which method helps find the best number of clusters in K-means?
Cross-validation
Sigmoid curve
Elbow method
Gradient descent
What does PCA help with?
Assigning labels
Finding data errors
Reducing data dimensions
Increasing features
In PCA, what are principal components?
Input vectors
Random guesses
New directions with most variance
Data errors
Which of these is not unsupervised learning?
PCA
K-means
Agglomerative clustering
Logistic regression
