Selecting a Form of ML CT-AI Practice Questions (Page 2)

Questions 6–10

1. 6expert · hard

   A financial firm wants to detect fraudulent transactions. They have a large dataset of transactions, but only a tiny fraction are labeled as fraudulent. The cost of missing a fraud is high, but the cost of false alarms is also significant. The team needs a model that can identify suspicious patterns without requiring extensive labeled data. Which approach is most appropriate?

   • AUse supervised regression to predict the fraud score
   • BUse supervised classification with the small labeled dataset only
   • CUse reinforcement learning to maximize the reward for catching fraud
   • DUse unsupervised anomaly detection to flag transactions that deviate from normal behavior
   Reveal Solution Discussion Select an answer first
2. 7expert · hard

   A logistics company wants to predict delivery delays. They have two datasets: (1) historical shipment records with labels indicating whether each shipment was delayed, and (2) real-time traffic data that is continuously updated. The company needs a model that can be retrained periodically as new labeled data arrives. Which approach best balances accuracy and maintainability?

   • ATrain a supervised classification model on the historical labeled data and retrain it periodically
   • BUse reinforcement learning to optimize delivery routes in real time
   • CUse unsupervised clustering on the traffic data to identify delay patterns
   • DWrite a set of static rules based on traffic conditions
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A streaming service wants to build a recommendation system that suggests movies to users. They have a large dataset of user ratings (e.g., 1-5 stars) for movies. The goal is to predict how a user would rate a movie they haven't seen. Which form of ML is most appropriate?

   • ASupervised classification to predict whether a user will like a movie
   • BUnsupervised clustering to group similar users
   • CSupervised regression to predict the rating value
   • DReinforcement learning to optimize recommendations over time
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A robotics company is developing a robot arm that must learn to pick up objects of varying shapes. The arm should improve its grip strategy through trial and error, receiving a positive signal when it successfully grasps an object and a negative signal when it drops it. Which form of ML best fits this requirement?

   • ADimensionality reduction to simplify the sensor data
   • BSupervised learning with a dataset of successful and failed grasps
   • CUnsupervised learning to cluster object shapes
   • DReinforcement learning with rewards for successful grasps
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   In supervised learning, what is the key difference between classification and regression tasks?

   • AClassification requires a neural network, while regression requires a decision tree.
   • BClassification predicts a discrete class label, while regression predicts a continuous numeric value.
   • CClassification is used for unlabeled data, while regression is used for labeled data.
   • DClassification predicts a continuous value, while regression predicts a discrete label.
   Reveal Solution Discussion Select an answer first