Verify Data Quality for Go/no-Go Decision to Conduct Data Preparation PMI-CPMAI Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   What is the primary purpose of identifying and evaluating bias in a dataset?

   • ATo ensure the model performs equally well across all groups
   • BTo increase the accuracy of the model on the training data
   • CTo reduce the size of the dataset
   • DTo make the data collection process faster
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   A data science team completed the final data quality assessment for a churn-prediction model. They found that 8% of the 'monthly_charges' values were missing and that the 'contract_type' column had inconsistent values ('Month-to-month' vs 'monthly'). The team imputed the missing values and standardized the contract_type values. What should the team do next?

   • ARe-run the entire data collection process to avoid any residual issues.
   • BAsk the business stakeholders if the imputation method is acceptable.
   • CProceed to model training because all identified issues have been resolved.
   • DDocument the issues, the imputation method, and the standardization steps in the data quality report.
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Which element should be included in the documentation of data quality findings?

   • AThe issues identified and the actions taken to address them
   • BThe personal opinions of the data scientists
   • CThe exact code used to train the final model
   • DA list of all possible machine learning algorithms
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   In a data representativeness assessment, what does 'under-representation' of a subgroup indicate?

   • AThe subgroup appears in the dataset less frequently than in the target population
   • BThe subgroup is perfectly balanced with other groups
   • CThe subgroup appears more frequently than in the target population
   • DThe subgroup has no missing values
   Reveal Solution Discussion Select an answer first
6. 15expert · hard

   A bank is developing a model to approve small business loans. The training data includes loan applications from the past 3 years, but the bank recently changed its underwriting policy to be more inclusive of businesses with thin credit files. The model will be used to evaluate new applications under the new policy. The team has a limited budget and cannot collect new data. What is the most appropriate action?

   • ATrain the model on historical data and restrict its use to applications that match the old policy criteria.
   • BDelay the go decision and explain that the historical data does not reflect the new underwriting policy.
   • CUse the historical data but adjust the model's decision threshold to be more inclusive.
   • DProceed with training on historical data because the model can learn the new policy from the data.
   Reveal Solution Discussion Select an answer first