Logistic Regression DATA-SCIENCE-FOUNDATIONS Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A logistic regression model for loan approval includes two predictors: 'credit score' and 'credit score squared'. The coefficient for the squared term is statistically significant. What is the most appropriate interpretation?

   • AMulticollinearity is present and the model should be simplified
   • BThe relationship between credit score and the log-odds of approval is non-linear
   • CThe model is invalid because logistic regression requires linearity
   • DThe coefficient for credit score is now meaningless
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Which metric is defined as the number of true positives divided by the sum of true positives and false positives?

   • AAccuracy
   • BPrecision
   • CF1-score
   • DRecall
   Reveal Solution Discussion Select an answer first
4. 3expert · hard

   A logistic regression model is trained on a dataset with 1,000 observations and 50 predictors. The model shows perfect separation on the training set (all predictions correct), but the coefficient estimates are extremely large. What is the most likely issue?

   • AThe model is fine; large coefficients indicate strong predictors
   • BMulticollinearity among predictors, inflating standard errors
   • CComplete separation, which causes MLE to produce unstable and inflated coefficients
   • DOverfitting due to too many predictors, but coefficients are still valid
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   Which of the following is a limitation of logistic regression?

   • AIt assumes a linear decision boundary between classes
   • BIt can only handle continuous predictor variables
   • CIt requires the outcome variable to have more than two categories
   • DIt cannot be used for prediction on new data
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A logistic regression model for credit risk includes highly correlated features such as 'number of late payments' and 'number of missed payments'. Which diagnostic issue is most likely?

   • AHeteroscedasticity, which violates the constant variance assumption
   • BNon-normality of residuals, which invalidates the model
   • CMulticollinearity, which can inflate standard errors of coefficients
   • DOutliers causing a non-linear decision boundary
   Reveal Solution Discussion Select an answer first