How Data, Algorithms, and Models Form AI Foundations AIE Practice Questions (Page 2)

Questions 6–10

1. 6expert · hard

   A retail chain wants to forecast product demand. They have two years of daily sales data. They train a model and achieve excellent performance on historical data, but when deployed, the forecasts are inaccurate. What is the most likely cause?

   • AThe model has overfit to historical patterns that do not generalize to future conditions
   • BThe algorithm is not powerful enough to learn sales patterns
   • CThe data is too clean and needs more noise
   • DThe model is underfitting because it has too many parameters
   Reveal Solution Discussion Select an answer first
2. 7expert · hard

   A financial firm wants to build a credit scoring model. They have a large dataset with many features. They are considering using a deep neural network versus a logistic regression model. The firm needs to explain decisions to regulators. What is the most important trade-off?

   • ADeep neural networks may have higher accuracy but lower interpretability compared to logistic regression
   • BDeep neural networks are easier to explain than logistic regression
   • CThe choice of algorithm does not affect model interpretability
   • DLogistic regression always has higher accuracy than deep neural networks
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A bank wants to approve or deny loan applications automatically. They have historical data on approved and denied loans. After training a decision tree on this data, what does the resulting model represent?

   • AThe decision tree algorithm's source code
   • BThe raw historical loan data stored in a database
   • CA list of all loan applications received in the past year
   • DA set of learned rules that map applicant features to loan decisions
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A manufacturing company wants to predict equipment failure before it happens. They have sensor data (temperature, vibration, pressure) collected over time, along with maintenance logs indicating failures. They train a random forest model. What does the model represent after training?

   • AA set of maintenance rules written by engineers
   • BA learned relationship between sensor readings and the likelihood of failure
   • CThe random forest algorithm's source code
   • DThe raw sensor data used for training
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A ride-sharing company wants to predict ride demand in different city zones. They have historical ride data, weather data, and event schedules. They plan to use a gradient boosting algorithm. What is the correct sequence of steps to build this predictive system?

   • ASelect a model, then find an algorithm that fits the model
   • BCollect and preprocess data, then use the algorithm to train a model, then use the model to predict demand
   • CTrain the model first, then collect data to validate it
   • DUse the algorithm to generate data, then build a model from that data
   Reveal Solution Discussion Select an answer first