NEW
Font size
WorksheetsMachine Learning Basics Worksheet
Total questions: 50
Worksheet time: 38mins
Aarush is a data scientist working on a project to categorize different types of fruits based on their characteristics. He needs to classify the fruits into various categories such as apples, bananas, and oranges. What is multiclass classification?
Classifying data into only two categories
Classifying data without labels
Classifying data into more than two categories
Reducing dimensionality of data
Which of the following is a type of Machine Learning?
Supervised Learning
Unsupervised Learning
Reinforcement Learning
All of the above
Which algorithm is used for classification problems?
K-Means
Linear Regression
Logistic Regression
PCA
What does a confusion matrix show?
Confusion in data
Errors in model code
Actual vs. predicted classifications
None of the above
Which of the following is a distance-based algorithm?
Naive Bayes
K-Nearest Neighbors (KNN)
Decision Tree
Logistic Regression
Which of the following is used for reducing the dimensionality of data?
KNN
Decision Tree
PCA
SVM
Which algorithm works best with linearly separable data?
Decision Tree
KNN
Support Vector Machine (SVM)
Naive Bayes
Aisha is training a machine learning model to predict house prices based on various features. However, she notices that the model performs exceptionally well on the training data but fails to predict prices accurately on new, unseen data. This situation is known as:
The model learns the noise in training data
The model generalizes well
The model is too simple
The data is clean
In a recent project, a team of data scientists was evaluating different models to predict customer churn. They were discussing various performance metrics to assess their models. During the discussion, Alisha mentioned several metrics, but Kavya pointed out one that is not typically used for performance evaluation. Which one is NOT a valid performance metric?
Accuracy
Recall
Mean Squared Error
Data Type
Unsupervised learning is used when:
Data is labeled
Output is known
No labels are available
Model training is not required
Saisha is looking for a new movie to watch on a streaming platform. The platform uses various techniques to recommend movies to its users. Which technique is used in recommendation systems?
Clustering
Regression
Collaborative Filtering
Classification
What is the goal of regression?
Group data
Reduce noise
Predict continuous values
Split datasets
Which of these is NOT a supervised learning algorithm?
Decision Tree
Linear Regression
K-Means
Naive Bayes
In Reinforcement Learning, the agent learns by:
Being told the correct answer
Using only unsupervised methods
Receiving rewards or penalties
Clustering data
Avni is working on a machine learning project and needs to improve her model's performance. She decides to use feature selection to:
Reduce model complexity
Increase noise
Reduce training data
Skip data preprocessing
What is an epoch in training a model?
One complete pass through the entire training dataset
A data preprocessing step
A type of model
A feature reduction method
Ensemble learning means:
Using one model at a time
Merging data
Combining multiple models for better accuracy
Clustering features
What is the output of a classification model?
Continuous value
Probability or class label
Unlabeled data
Graph
Which algorithm is best for spam detection?
K-Means
Naive Bayes
PCA
SVM
Which of the following is NOT a supervised learning algorithm?
Linear Regression
Decision Tree
K-Means Clustering
Support Vector Machine
Shreya is trying to decide which machine learning model to use for her project. She is considering various options and wants to know the drawbacks of using Decision Trees. Which of the following is a disadvantage of Decision Trees?
Easy to interpret
Handles both numerical and categorical data
High accuracy
Prone to overfitting
What is the purpose of the training dataset in supervised learning?
To predict unseen data
To adjust model parameters
To test the accuracy
To reduce noise
In Support Vector Machines (SVM), what is the purpose of the kernel function?
Reduce training time
Handle linear data only
Transform data into higher dimensions
Normalize the input data
What does the K in K-Nearest Neighbors represent?
Number of features
Number of iterations
Number of classes
Number of nearest neighbors
In supervised learning, the model is trained using:
Only input data
Only output data
Input-output pairs (labeled data)
Unlabelled data
What is the main goal of ensemble learning?
To reduce dataset size
To increase model complexity
To combine multiple models to improve performance
To reduce training time
Which of the following is an ensemble learning method?
Linear Regression
Decision Tree
Random Forest
K-Nearest Neighbor
In Boosting, models are trained:
In parallel
On the same data repeatedly
Sequentially, each focusing on previous errors
Using unsupervised data
What is the purpose of pruning in decision trees?
To increase tree depth
To reduce overfitting
To add more branches
To increase accuracy on training data
Post-pruning is performed:
During tree building
After the entire tree is built
Before splitting the nodes
On training data only
Overfitting in decision trees usually happens when:
The tree is too shallow
The tree is pruned too early
The tree grows too deep without pruning
The training data is small
What is the basic unit of an artificial neural network?
Weight
Neuron (Node)
Layer
Activation
In a neural network, the function that introduces non-linearity is called:
Cost Function
Gradient Function
Activation Function
Weight Function
Which of the following is NOT a commonly used activation function?
Sigmoid
ReLU
Tanh
Euclidean
What does a weight in a neural network represent?
The output of a neuron
The amount of learning
The strength of the connection between neurons
The number of hidden layers
The process of adjusting weights in a neural network is called:
Forward Propagation
Normalization
Backpropagation
Pruning
Which algorithm is typically used to train an artificial neural network?
Naive Bayes
K-Means
Gradient Descent
Apriori
A network with more than one hidden layer is called a:
Convolutional Neural Network
Deep Neural Network
Simple Perceptron
Feedforward Network
What is the role of the bias term in a neuron?
To set the weight value to zero
To reduce the cost function
To allow shifting of the activation function
To deactivate the neuron
Which of the following is true about the learning rate in neural networks?
A. A very high learning rate guarantees faster convergence
B. A low learning rate may slow training but lead to better accuracy
C. Learning rate does not affect backpropagation
D. Learning rate increases with each epoch
What is the role of the loss function in a neural network?
Determines number of neurons
Measures prediction accuracy
Computes the error to be minimized
Selects activation functions
What does ReLU stand for in activation functions?
Rectified Learning Unit
Rectified Linear Unit
Recursive Learning Unit
Regression Line Unit
Which of the following is NOT a type of layer in a neural network?
Input Layer
Output Layer
Expansion Layer
Hidden Layer
Which function is used in binary classification problems in the output layer of a neural network?
ReLU
Tanh
Softmax
Sigmoid
Which of the following is TRUE about a Perceptron?
It can solve non-linear problems
It has multiple hidden layers
It is the simplest form of a neural network
It requires pooling layers
What technique does Random Forest use to improve accuracy and reduce overfitting?
Feature scaling
Bagging (Bootstrap Aggregation)
Dimensionality reduction
Pruning
In SVM, what are support vectors?
The input features
Data points closest to the decision boundary
Outliers
Parameters of the kernel
Which of the following is a key advantage of Random Forest?
Requires a lot of pre-processing
Performs poorly on large datasets
Handles missing data well
Cannot be used for classification tasks
Random Forest is an ensemble of:
Logistic regression models
Support vector machines
Decision trees
Naive Bayes classifiers
What is the main goal of unsupervised learning?
Predict output values
Train on labeled data
Find patterns or structure in unlabelled data
Maximize accuracy
