Supervised Machine Learning Overview Answers 1. What does Mean Absolute Error measure in the context of model predictions? A. The average difference without considering the direction of errors B. The squared differences between predicted and actual values C. The total number of predictions made by a model D. The maximum error in a single prediction Answer: The average difference without considering the direction of errors (A) Mean Absolute Error calculates the average absolute difference between predicted and actual values, providing insight into overall prediction accuracy. 2. What is the primary purpose of cross-validation in model selection? A. To eliminate the need for hyperparameter tuning B. To evaluate the model's performance on unseen data C. To find the best training set for the model D. To reduce the size of the dataset by removing noise Answer: To evaluate the model's performance on unseen data (B) Cross-validation is primarily used to estimate how well a model will perform on unseen data, ensuring a more reliable model choice. 3. Which situation best describes overfitting in a model? A. The model accurately predicts both training and new data B. The model performs exceptionally on training data but poorly on new data C. The model is too simplistic and cannot make accurate predictions D. The model fails to capture underlying data patterns Answer: The model performs exceptionally on training data but poorly on new data (B) Overfitting occurs when a model learns the training data too well, including noise and outliers, leading to poor performance on unseen data. 4. In which application would supervised learning most likely be used? A. Clustering similar items together B. Identifying spam emails based on their content C. Creating new data points based on existing ones D. Predicting stock market trends without labeled data Answer: Identifying spam emails based on their content (B) Supervised learning is ideal for applications like spam filtering, where there is labeled data to train the model. 5. What does hyperparameter tuning aim to achieve in machine learning models? A. To reduce the size of the training dataset B. To automate the training process completely C. To eliminate the need for cross-validation D. To find the optimal values for parameters set prior to training Answer: To find the optimal values for parameters set prior to training (D) Hyperparameter tuning focuses on optimizing parameters that are not learned during the training process to enhance model performance. 6. What is the primary goal of a classification task in supervised machine learning? A. To minimize the error in predictions. B. To predict continuous output variables. C. To rank data points based on input features. D. To predict a categorical output variable. Answer: To predict a categorical output variable. (D) Classification tasks are specifically designed to predict categorical outputs, such as distinguishing between different classes. 7. Which algorithm is best suited for high-dimensional data in supervised machine learning? A. Linear Regression B. Support Vector Machines (SVM) C. Decision Trees D. Naive Bayes Answer: Support Vector Machines (SVM) (B) Support Vector Machines (SVM) are particularly effective for handling highdimensional data. 8. In the context of regression tasks, which metric measures the average difference between predicted and actual values? A. Precision B. Root Mean Squared Error (RMSE) C. Recall D. F1-score Answer: Root Mean Squared Error (RMSE) (B) Root Mean Squared Error (RMSE) is specifically designed to measure the average error in regression predictions. 9. What does logistic regression output in supervised machine learning? A. A continuous value representing a target variable. B. A linear equation describing variable relationships. C. A probability between 0 and 1 indicating class membership. D. A discrete classification based on input features. Answer: A probability between 0 and 1 indicating class membership. (C) Logistic regression is focused on predicting probabilities that a data point belongs to a specific class. 10. Which performance metric is crucial to consider when false positives are costly? A. F1-score B. Recall C. Accuracy D. Precision Answer: Precision (D) Precision assesses the accuracy of positive predictions, which is vital when false positives can have significant consequences. 11. What is a notable weakness of decision trees in supervised machine learning? A. They are prone to overfitting. B. They are overly complex and difficult to interpret. C. They cannot handle imbalanced datasets. D. They cannot be used for classification tasks. Answer: They are prone to overfitting. (A) Decision trees can easily overfit the training data, particularly if they are excessively complex. 12. Which of the following algorithms assumes features are conditionally independent given the class? A. Support Vector Machines (SVM) B. Logistic Regression C. Naive Bayes D. Polynomial Regression Answer: Naive Bayes (C) Naive Bayes relies on the assumption of independence among features, based on Bayes' theorem. 13. Which metric is considered a balanced measure combining precision and recall? A. Recall B. Accuracy C. F1-score D. Root Mean Squared Error (RMSE) Answer: F1-score (C) The F1-score is the harmonic mean of precision and recall, providing a balance between the two.