Objective 2.5 Address Business Risks, Ethical Concerns, and Related Concepts in Data Exploration/feature Engineering AIP-210 Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   Which metric is commonly used to evaluate fairness by comparing the true positive rates across different groups?

   • APredictive parity
   • BEqualized odds
   • CEqual opportunity
   • DDemographic parity
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A data science team has identified a data quality issue in a feature used for a predictive maintenance model. The issue could lead to incorrect predictions and costly downtime for clients. The team needs to inform the sales team, who are not technical, about the potential impact. What is the most effective communication approach?

   • ADescribe the technical details of the data quality issue, such as missing values and imputation methods.
   • BProvide the sales team with the full data lineage report and ask them to read it.
   • CExplain the potential financial impact on clients in terms of unplanned downtime and maintenance costs.
   • DReassure the sales team that the model is still 'good enough' and no action is needed.
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   Why is it important to communicate ethical risks to stakeholders?

   • ATo increase the technical complexity of the project
   • BTo satisfy regulatory requirements only
   • CTo shift responsibility to the stakeholders
   • DTo ensure stakeholders can make informed decisions about the AI system
   Reveal Solution Discussion Select an answer first
5. 24application · medium

   A data science team is working on a model for a retail company. They receive a new data source from a third-party vendor. Before using this data for feature engineering, what is the most important action to take?

   • AImmediately integrate the data into the existing feature pipeline.
   • BTrain a separate model using only the new data to test its predictive power.
   • CDocument the source, the date of acquisition, and any known limitations of the data.
   • DRemove all personally identifiable information from the data before use.
   Reveal Solution Discussion Select an answer first
6. 25foundation · easy

   What is the purpose of using fairness metrics in machine learning?

   • ATo detect and quantify bias in model predictions across different groups
   • BTo reduce the training time of the model
   • CTo ensure the model is interpretable
   • DTo increase the overall accuracy of the model
   Reveal Solution Discussion Select an answer first