AI / MACHINE LEARNING FEATURE MODELLING — Learn free

Self-check

Questions 1–10 of 10

1. In AI / Machine Learning Feature Modelling, what is a feature?
Question 1A. A screen in a machine learning applicationB. A measurable variable used as an input to a machine learning modelC. A column in a dimension tableD. A rule in a security policyE. A version of a trained model
2. What is the primary goal of Feature Modelling?
Question 2A. To let every data scientist build their own version of each metricB. To replace dimensional modelling in the warehouseC. To provide a standardised, governed framework so features are defined once and reused consistentlyD. To store raw transactional data directly in ML algorithmsE. To eliminate the need for data stewards
3. What is a Feature Store?
Question 3A. A marketplace where third-party models are soldB. A backup of a data warehouseC. A dashboard showing model accuracyD. A centralised repository where engineered features are stored, versioned and made available for training and inferenceE. A list of security roles for data scientists
4. Which of the following must a Feature Definition include?
Question 4A. A clear name, calculation logic, data sources, refresh frequency and ownershipB. The model's hyperparametersC. The GPU used for trainingD. A star schema diagramE. The list of end users of the model
5. How is Repayment_DelinquencyRate defined in the banking example from I Am Datapedia!?
Question 5A. The number of prior loans a customer has takenB. The percentage of approvals by each officer over six monthsC. The customer's credit score from the bureauD. The ratio of late repayments over the past 12 monthsE. The average loan amount per branch
6. Which feature in the loan approval example is sourced from credit bureau data and refreshed monthly?
Question 6A. Loan_HistoryCountB. Officer_ApprovalRateC. Customer_CreditScoreD. Repayment_DelinquencyRateE. Customer_IncomeStability
7. How does Feature Modelling differ from dimensional modelling and metadata modelling?
Question 7A. It is focused on analytical reportingB. It describes data assets in generalC. It defines physical storage formatsD. It is only used for unstructured dataE. It is explicitly model-input-centric, sitting between data engineering and ML operations
8. Why do data scientists retrieve features from the Feature Store rather than recalculating them?
Question 8A. To ensure consistency between training and production environments and support explainable AIB. Because recalculation is forbidden by GDPRC. Because the Feature Store is faster than any databaseD. To avoid paying for computeE. Because raw data is never stored anywhere else
9. Which is described as a drawback of a governed feature catalogue?
Question 9A. It increases duplication of featuresB. Excessive governance can slow experimentation in early-stage innovationC. It makes models less explainableD. It weakens regulatory complianceE. It prevents reuse across models
10. Which of the following is NOT named as a use case that could reuse the same engineered features?
Question 10A. Risk scoringB. Fraud detectionC. Churn predictionD. Payroll processingE. Loan approval