Objective 3.4 Evaluate the Model AIP-210 Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   In the bias-variance tradeoff, what does high variance typically indicate about a model?

   • AThe model is too simple and underfits the data
   • BThe model has low error on both training and validation sets
   • CThe model is sensitive to small changes in the training data
   • DThe model performs consistently across different datasets
   Reveal Solution Discussion Select an answer first
2. 2expert · hard

   A medical diagnostic model has an AUC of 0.95, but at the default decision threshold, the precision is 0.5 and recall is 0.9. The clinical team wants to minimize false positives to avoid unnecessary procedures. Which action is most appropriate?

   • AIncrease the decision threshold to improve precision at the cost of recall
   • BDecrease the decision threshold to improve recall
   • CUse the AUC directly to make decisions
   • DReject the model because precision is low
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   After fitting a linear regression model, a data scientist plots the residuals against the predicted values. The plot shows a clear funnel shape, with residuals spreading wider as predicted values increase. What does this indicate, and what action should be taken?

   • AHeteroscedasticity; consider applying a log transformation to the target variable
   • BAutocorrelation; use time series methods
   • CNon-linearity; add polynomial features to the model
   • DMulticollinearity; remove correlated features
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   In residual analysis for a regression model, what pattern in a plot of residuals versus fitted values suggests a violation of the homoscedasticity assumption?

   • AResiduals are all positive and increase linearly with fitted values
   • BResiduals follow a perfect straight line
   • CResiduals are randomly scattered around zero with constant spread
   • DResiduals form a funnel shape, with spread increasing as fitted values increase
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   A data scientist is using a complex model that achieves low training error but high validation error. The scientist considers two options: (1) increase the amount of training data, or (2) reduce model complexity. Which statement best describes the tradeoff?

   • AIncreasing data always reduces bias, while reducing complexity always reduces variance
   • BIncreasing data can reduce variance without increasing bias, but may be costly or infeasible; reducing complexity increases bias but reduces variance
   • CNeither option will help because the model is already overfit
   • DBoth options are equivalent in their effect on the bias-variance tradeoff
   Reveal Solution Discussion Select an answer first