Identify Common Machine Learning Techniques AI-900 Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A logistics company wants to forecast the delivery time for each package based on distance, traffic, and weather. They also want to identify packages that are likely to be delayed. Which combination of techniques should they use?

   • ADeep learning to predict delivery time, and clustering to flag delays.
   • BRegression to predict delivery time, and classification to flag delayed packages.
   • CClassification to predict delivery time, and clustering to identify delayed packages.
   • DClustering to group packages by delivery time, and regression to flag delays.
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A retail company wants to predict the daily sales amount for each of its 500 stores for the next 30 days. The data includes historical sales, store location, local holidays, and weather conditions. The company also wants to automatically flag stores whose predicted sales deviate significantly from actual sales as 'unusual' for further review. Which combination of machine learning tasks should the company use?

   • AUse a transformer model to generate sales text summaries, then use regression to detect anomalies.
   • BUse clustering to group stores by sales pattern, then use classification to label each day as 'normal' or 'anomaly'.
   • CUse classification to predict sales as 'high' or 'low', then use regression to flag unusual stores.
   • DUse regression to forecast the sales amount, then use anomaly detection (clustering-based) to flag deviations.
   Reveal Solution Discussion Select an answer first
4. 3expert · hard

   A retail chain wants to segment its stores based on sales performance, customer footfall, and product mix. They do not have predefined segments. They also want to identify stores that are underperforming compared to their segment. Which approach should they use?

   • AUse clustering to segment stores, and then use a classification model to identify underperforming stores within each segment.
   • BUse classification to segment stores, and clustering to identify underperformers.
   • CUse deep learning to segment stores and identify underperformers in one model.
   • DUse regression to segment stores, and clustering to identify underperformers.
   Reveal Solution Discussion Select an answer first
5. 4expert · hard

   A real estate company wants to predict house prices (in dollars) and also categorize houses into 'underpriced', 'fairly priced', or 'overpriced' based on the prediction error. They have historical sales data with features like square footage, location, and number of bedrooms. They need to automate both tasks. Which approach should they take?

   • AUse a transformer to generate a text description of each house.
   • BTrain a regression model to predict the price, then use the prediction error to classify into the three categories.
   • CUse clustering to group houses by price, then label each group.
   • DTrain a classification model directly to predict the three categories.
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   Which mechanism is a key feature of the Transformer architecture?

   • ADecision trees
   • BSelf-attention
   • CRecurrent connections
   • DConvolutional filters
   Reveal Solution Discussion Select an answer first