Download Machine Learning Expert Exam and more Exams Technology in PDF only on Docsity!
Question 1 Question: What is the main difference between supervised and unsupervised learning? A. Supervised learning uses labeled data, unsupervised learning does not B. Supervised learning is faster than unsupervised learning C. Unsupervised learning requires labeled data, supervised does not D. Unsupervised learning can only be used for regression Correct: A Explanation: Supervised learning utilizes labeled data to train the model, whereas unsupervised learning works with unlabeled data to find patterns or clusters. Question 2 Question: Which of the following is NOT a type of machine learning? A. Reinforcement Learning B. Supervised Learning C. Unsupervised Learning
D. Deterministic Learning Correct: D Explanation: Deterministic learning is not a standard category of machine learning; the main types are supervised, unsupervised, semi- supervised, and reinforcement learning. Question 3 Question: In traditional programming, you provide data and rules to produce answers. In machine learning, you provide data and answers to produce what? A. Features B. Algorithms C. Rules D. Labels Correct: C Explanation: In ML, data and labels (answers) are used to learn the rules (the model).
C. Logistic Regression D. Apriori Algorithm Correct: C Explanation: Logistic regression is widely used for classification tasks. Question 6 Question: What is the dimension of a matrix with 4 rows and 3 columns? A. 3x B. 4x C. 7 D. 12 Correct: B Explanation: Matrix dimensions are given as rows × columns, so 4x3. Question 7 Question: What is the eigenvalue of a matrix?
A. A scalar that stretches the eigenvector B. A vector perpendicular to the matrix C. The sum of the matrix elements D. The determinant of the matrix Correct: A Explanation: Eigenvalues are scalars that, when multiplied with their corresponding eigenvectors, result in the same direction as the original matrix transformation. Question 8 Question: Which of the following is a discrete probability distribution? A. Normal distribution B. Bernoulli distribution C. Exponential distribution D. Uniform distribution (continuous) Correct: B Explanation: The Bernoulli distribution is discrete, representing outcomes of a binary experiment.
D. Mode Correct: C Explanation: Variance quantifies how much data points differ from the mean. Question 11 Question: What does gradient descent optimize in machine learning models? A. Loss function B. Data size C. Output labels D. Feature selection Correct: A Explanation: Gradient descent minimizes the loss function, improving model accuracy.
Question 12 Question: What is the primary difference between stochastic and batch gradient descent? A. Stochastic uses one sample at a time; batch uses all samples B. Stochastic uses all samples; batch uses one sample C. Stochastic is slower than batch D. Batch can only be used for deep learning Correct: A Explanation: Stochastic gradient descent updates parameters using a single sample, while batch gradient descent uses the entire dataset. Question 13 Question: Which optimization algorithm adapts learning rates for each parameter? A. Newton’s Method B. Adam C. Gradient Descent D. Least Squares
continuous value? A. Logistic Regression B. Decision Tree Classification C. Linear Regression D. Naive Bayes Correct: C Explanation: Linear regression predicts continuous outcomes. Question 16 Question: What is the main purpose of regularization in regression? A. Increase model complexity B. Reduce overfitting C. Improve training time D. Increase variance Correct: B Explanation: Regularization penalizes large coefficients, reducing overfitting.
Question 17 Question: Which regularization technique can shrink some coefficients to zero? A. L1 (Lasso) B. L2 (Ridge) C. Both L1 and L D. None Correct: A Explanation: L1 (Lasso) regularization can shrink coefficients exactly to zero, aiding feature selection. Question 18 Question: What is the output of a logistic regression model? A. Probability between 0 and 1 B. Integer values C. Continuous values
problems? A. SVM with linear kernel B. SVM with RBF kernel C. Linear Regression D. Logistic Regression Correct: B Explanation: SVM with RBF kernel can capture non-linear relationships. Question 21 Question: What is the k in k-Nearest Neighbors (KNN) algorithm? A. Number of clusters B. Number of neighbors to consider C. Number of decision trees D. Learning rate Correct: B Explanation: In KNN, k indicates how many nearest neighbors are considered for voting or averaging.
Question 22 Question: Which classifier assumes independence between features? A. Logistic Regression B. Naive Bayes C. Decision Tree D. SVM Correct: B Explanation: Naive Bayes assumes features are independent given the class. Question 23 Question: Which metric is suitable for imbalanced classification problems? A. Accuracy B. Precision, Recall, F1-Score C. Mean Squared Error
A. Testing model speed B. Evaluating model generalization C. Increasing data size D. Reducing variance Correct: B Explanation: Cross-validation estimates how well a model generalizes to unseen data. Question 26 Question: What does grid search help with? A. Model selection and hyperparameter tuning B. Data preprocessing C. Feature scaling D. Model deployment Correct: A Explanation: Grid search exhaustively searches through a specified subset of hyperparameters to optimize model performance.
Question 27 Question: What is the main goal of clustering? A. Predicting future values B. Grouping similar data points C. Minimizing loss D. Maximizing variance Correct: B Explanation: Clustering groups data points based on similarity without using labeled data. Question 28 Question: Which algorithm is NOT used for clustering? A. K-Means B. DBSCAN C. Hierarchical Clustering D. Logistic Regression
A. Elbow Method B. ROC Curve C. Cross-Entropy D. PCA Correct: A Explanation: The Elbow Method helps identify the optimal number of clusters by plotting the sum of squared distances. Question 31 Question: What does Silhouette analysis measure in clustering? A. Accuracy B. Cluster cohesion and separation C. Model overfitting D. Learning rate Correct: B Explanation: Silhouette analysis quantifies how similar a data point is to its own cluster compared to other clusters.
Question 32 Question: What is the main goal of dimensionality reduction? A. Increase the number of features B. Reduce the number of features while preserving information C. Increase data variance D. Reduce the number of samples Correct: B Explanation: Dimensionality reduction aims to reduce the feature space while retaining as much information as possible. Question 33 Question: Which technique is commonly used for dimensionality reduction? A. Principal Component Analysis (PCA) B. KNN C. Random Forest
