Which measure cannot be used to evaluate the performance of a logistic regression model?

Which measure cannot be used to evaluate the performance of a logistic regression model?

• A. Accuracy
• B. R-squared
• C. Confusion matrix
• D. AUC-ROC

Answer: (B)

R-squared is a statistical measure that indicates how well data fits a regression model, commonly used in linear regression. It represents the proportion of variance for the dependent variable that is explained by the independent variables in the model. However, its application in logistic regression is inappropriate because logistic regression is designed for binary outcomes, where the response variable is categorical rather than continuous. In logistic regression, the predicted values are probabilities that classify observations into one of the two possible classes rather than predicting a continuous outcome. Consequently, measures like accuracy, confusion matrix, and AUC-ROC are more suitable for evaluating the performance of logistic regression models. Accuracy measures the proportion of correctly classified instances among the total instances, while the confusion matrix provides a comprehensive view of the model's performance by showing true positives, false positives, true negatives, and false negatives. AUC-ROC evaluates the model's ability to distinguish between the two classes across different threshold values, summarizing the trade-off between sensitivity and specificity. These metrics provide relevant insights into how well the logistic regression model performs in classifying data points into their respective categories.

R-squared is a statistical measure that indicates how well data fits a regression model, commonly used in linear regression. It represents the proportion of variance for the dependent variable that is explained by the independent variables in the model. However, its application in logistic regression is inappropriate because logistic regression is designed for binary outcomes, where the response variable is categorical rather than continuous.

In logistic regression, the predicted values are probabilities that classify observations into one of the two possible classes rather than predicting a continuous outcome. Consequently, measures like accuracy, confusion matrix, and AUC-ROC are more suitable for evaluating the performance of logistic regression models.

Accuracy measures the proportion of correctly classified instances among the total instances, while the confusion matrix provides a comprehensive view of the model's performance by showing true positives, false positives, true negatives, and false negatives. AUC-ROC evaluates the model's ability to distinguish between the two classes across different threshold values, summarizing the trade-off between sensitivity and specificity. These metrics provide relevant insights into how well the logistic regression model performs in classifying data points into their respective categories.