Data Quality and Its Effect on the ML Model CT-AI Practice Questions (Page 5)

Questions 21–25

1. 21expert · hard

   A city government is building a model to allocate resources for public services. The training data is collected from a mobile app that is primarily used by younger, tech-savvy residents. The model performs well on this group but poorly on older residents who do not use the app. What is the ROOT CAUSE of this performance gap?

   • AThe data has a timeliness issue because the app was introduced recently.
   • BThe data has a consistency issue because the app records data differently than other sources.
   • CThe data has a completeness issue because older residents are underrepresented.
   • DThe data is not representative of the entire population, leading to biased model predictions.
   Reveal Solution Discussion Select an answer first
2. 22expert · hard

   A financial institution is building a model to detect fraudulent transactions. The dataset has a 'transaction_amount' column with a few extreme values (e.g., $1,000,000) that are legitimate (large wire transfers). The team is debating whether to cap these values. Which consideration is MOST important in this decision?

   • ACapping is always necessary to prevent outliers from skewing the model.
   • BCapping will reduce the model's ability to detect fraud on large legitimate transactions.
   • CCapping is the only way to handle outliers in a fraud-detection model.
   • DCapping will increase the model's accuracy because it reduces variance.
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   In the context of machine learning, what does 'data quality' primarily refer to?

   • AThe speed at which data can be processed by the ML pipeline
   • BThe degree to which data meets the requirements of the intended ML use case
   • CThe amount of data available for training the model
   • DThe security measures protecting the data from unauthorized access
   Reveal Solution Discussion Select an answer first
5. 24application · medium

   A logistics company is building a model to predict delivery times. The dataset includes a 'distance' field, but the units are inconsistent: some records use kilometers, others use miles. What data quality dimension is violated, and what is the BEST action?

   • AConsistency — convert all values to a single unit (e.g., kilometers) before training.
   • BAccuracy — the values are factually incorrect and should be re-measured.
   • CCompleteness — the missing unit information should be imputed.
   • DTimeliness — the data is outdated and should be refreshed.
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A retail company is building a churn-prediction model. The 'customer_since' field contains dates, but 15% of the records have dates in the future due to a system bug. The team must decide how to handle this before training. Which data quality dimension is being violated, and what is the MOST appropriate action?

   • AValidity — reject or correct the future dates because they do not conform to the expected range of the attribute.
   • BTimeliness — drop the records because the data is too old to be useful.
   • CConsistency — convert all dates to a single format to resolve the issue.
   • DCompleteness — impute the future dates with the median date because the field is partially filled.
   Reveal Solution Discussion Select an answer first