Q.1 What is the main goal of Machine Learning?
To write explicit programs
To learn patterns from data
To replace computer hardware
To design operating systems
Explanation - Machine learning focuses on enabling systems to automatically learn and improve from experience without being explicitly programmed.
Correct answer is: To learn patterns from data
Q.2 Which of the following is a type of Machine Learning?
Supervised Learning
Interpreted Learning
Scripted Learning
Manual Learning
Explanation - Supervised learning is a common type of machine learning where models are trained on labeled data.
Correct answer is: Supervised Learning
Q.3 In supervised learning, what do the training data include?
Only input data
Only output data
Input-output pairs
Random numbers
Explanation - Supervised learning uses datasets containing both input features and their corresponding correct outputs.
Correct answer is: Input-output pairs
Q.4 Which of the following is an example of unsupervised learning?
Spam email classification
Handwriting recognition
Customer segmentation
Stock price prediction
Explanation - Unsupervised learning deals with unlabeled data. Customer segmentation groups customers without predefined labels.
Correct answer is: Customer segmentation
Q.5 Which algorithm is commonly used for classification problems?
Linear Regression
K-Means
Decision Tree
PCA
Explanation - Decision Trees are widely used for classification tasks because they can handle categorical and numerical data effectively.
Correct answer is: Decision Tree
Q.6 What does 'overfitting' mean in machine learning?
Model performs well on training but poorly on test data
Model ignores training data
Model underfits training data
Model perfectly generalizes
Explanation - Overfitting occurs when a model memorizes training data instead of learning to generalize to unseen data.
Correct answer is: Model performs well on training but poorly on test data
Q.7 Which metric is commonly used to evaluate classification models?
Accuracy
Mean Squared Error
Variance
Entropy
Explanation - Accuracy measures the proportion of correct predictions among total predictions, making it suitable for classification tasks.
Correct answer is: Accuracy
Q.8 What type of learning is reinforcement learning?
Learning from labeled data
Learning from rewards and punishments
Learning without data
Learning from memory
Explanation - Reinforcement learning teaches agents to make sequences of decisions by rewarding good actions and penalizing bad ones.
Correct answer is: Learning from rewards and punishments
Q.9 Which of the following is an example of regression?
Predicting stock prices
Classifying spam emails
Grouping customers
Recommending movies
Explanation - Regression involves predicting continuous values, such as stock market prices.
Correct answer is: Predicting stock prices
Q.10 Which of the following best describes 'features' in machine learning?
The outcomes to predict
Input variables used for prediction
The learning rate
The reward signals
Explanation - Features are the independent variables used as inputs to the machine learning model.
Correct answer is: Input variables used for prediction
Q.11 What does a confusion matrix evaluate?
Model architecture
Training speed
Classification performance
Memory usage
Explanation - A confusion matrix provides detailed insight into classification results, showing true positives, false positives, etc.
Correct answer is: Classification performance
Q.12 Which algorithm is best known for clustering?
K-Means
Linear Regression
Naive Bayes
Logistic Regression
Explanation - K-Means is a popular unsupervised clustering algorithm that groups data into k clusters.
Correct answer is: K-Means
Q.13 Which machine learning model is inspired by the human brain?
Decision Trees
Neural Networks
Linear Regression
Support Vector Machines
Explanation - Neural networks are inspired by the structure and functioning of the human brain neurons.
Correct answer is: Neural Networks
Q.14 What is 'training data' in machine learning?
Data used to test model accuracy
Data used to tune hyperparameters
Data used to teach the model
Random unused data
Explanation - Training data is used by algorithms to learn patterns before being evaluated on test data.
Correct answer is: Data used to teach the model
Q.15 Which loss function is commonly used for regression?
Cross-Entropy Loss
Mean Squared Error
Hinge Loss
Log Loss
Explanation - Mean Squared Error is widely used in regression because it penalizes large errors heavily.
Correct answer is: Mean Squared Error
Q.16 What is 'dimensionality reduction' in machine learning?
Increasing features
Reducing number of features
Scaling output
Adding noise to data
Explanation - Dimensionality reduction reduces the number of input variables while retaining essential information.
Correct answer is: Reducing number of features
Q.17 Which of the following is a supervised learning algorithm?
K-Means
Linear Regression
Apriori Algorithm
DBSCAN
Explanation - Linear regression is a supervised algorithm as it uses labeled data for training.
Correct answer is: Linear Regression
Q.18 Which of these is NOT a type of machine learning?
Supervised Learning
Unsupervised Learning
Reinforcement Learning
Embedded Learning
Explanation - Embedded Learning is not a recognized category; the main types are supervised, unsupervised, and reinforcement learning.
Correct answer is: Embedded Learning
Q.19 What does 'bias' in a model represent?
Systematic error in predictions
Model variance
Random noise
Training speed
Explanation - Bias is the error from erroneous assumptions in the model that may underfit data.
Correct answer is: Systematic error in predictions
Q.20 What is 'variance' in machine learning?
Error due to bias
Error due to noise
Sensitivity to training data fluctuations
Irrelevant features
Explanation - Variance measures how much the model’s predictions change with different training datasets.
Correct answer is: Sensitivity to training data fluctuations
Q.21 Which algorithm is most suitable for spam detection?
Naive Bayes
K-Means
PCA
KNN
Explanation - Naive Bayes is commonly used for spam filtering due to its probabilistic approach and efficiency.
Correct answer is: Naive Bayes
Q.22 What is a 'label' in supervised learning?
The output value to predict
An input variable
A hidden feature
A clustering category
Explanation - Labels are the actual output values in supervised learning that the model tries to predict.
Correct answer is: The output value to predict
Q.23 Which method is used to prevent overfitting?
Cross-validation
Data ignoring
Adding noise
Underfitting
Explanation - Cross-validation helps assess model performance on unseen data, reducing overfitting risk.
Correct answer is: Cross-validation
Q.24 Which of the following is NOT an evaluation metric for classification?
Precision
Recall
F1-Score
Mean Squared Error
Explanation - Mean Squared Error is used for regression, not classification. Precision, Recall, and F1 are for classification.
Correct answer is: Mean Squared Error
Q.25 Which step comes first in a typical machine learning workflow?
Model training
Data preprocessing
Hyperparameter tuning
Model deployment
Explanation - Data preprocessing is essential before training as it ensures the data is clean and suitable for modeling.
Correct answer is: Data preprocessing