Introduction to Logistic Regression Answers 1. What is the primary purpose of logistic regression? A. To determine the correlation between multiple features B. To predict continuous outcomes C. To classify data points into multiple categories D. To model probabilities for binary classification problems Answer: To model probabilities for binary classification problems (D) Logistic regression is specifically designed for binary classification tasks. 2. Which function is central to the mechanics of logistic regression? A. Logistic function (sigmoid function) B. Linear function C. Quadratic function D. Exponential function Answer: Logistic function (sigmoid function) (A) The logistic function, or sigmoid function, is crucial in transforming linear outcomes into probabilities. 3. What does a positive coefficient in the logistic regression model indicate? A. Constant probability regardless of feature value B. Increased probability of a positive outcome with higher feature value C. No relationship between feature and outcome D. Decreased probability of a positive outcome Answer: Increased probability of a positive outcome with higher feature value (B) A positive coefficient implies that as the feature value increases, the probability of the specified outcome also increases. 4. What method is commonly used for estimating parameters in logistic regression? A. Least absolute deviations B. Gradient ascent C. Mean square error minimization D. Maximum likelihood estimation (MLE) Answer: Maximum likelihood estimation (MLE) (D) Maximum likelihood estimation is the typical approach for estimating parameters in logistic regression. 5. What defines the decision boundary in logistic regression? A. The point where probability equals 0 B. The threshold where predicted probabilities meet 0.75 C. The line separating classes where probability is 0.5 D. The area where all predicted outcomes are equal Answer: The line separating classes where probability is 0.5 (C) The decision boundary is established at the point where the predicted probability is 0.5. 6. Which metric is particularly useful for evaluating the performance of a model on imbalanced datasets? A. Recall B. Accuracy C. Log-loss D. F1-score Answer: F1-score (D) The F1-score balances precision and recall, making it effective in scenarios with class imbalances. 7. What does the recall metric measure in a classification context? A. The ratio of correctly predicted positive cases to total predicted positive cases B. The ratio of total positives predicted to the total dataset C. The ratio of correctly predicted positive cases to total actual positive cases D. The ratio of predicted positives to total negatives Answer: The ratio of correctly predicted positive cases to total actual positive cases (C) Recall indicates how well the model identifies actual positive cases among all actual positives. 8. Which of the following is a notable advantage of logistic regression? A. It can classify multiple classes simultaneously B. It is complex and requires extensive tuning C. It is simple to understand and implement D. It guarantees perfect accuracy in predictions Answer: It is simple to understand and implement (C) One of the main benefits of logistic regression is its simplicity, making it easy to implement and understand.