2.1 Recommend AI-powered Data Analytic Techniques to Meet Specific Needs or Answer Specific Questions 350-201 Practice Questions (Page 9)

Questions 41–45

1. 41foundation · easy

   What is the purpose of feature engineering in data preprocessing for AI?

   • ATo combine multiple models into one
   • BTo create new features that better represent the underlying patterns in the data
   • CTo label the data for supervised learning
   • DTo remove all features that are not numeric
   Reveal Solution Discussion Select an answer first
2. 42foundation · easy

   Which task would most likely use NLP techniques?

   • APredicting CPU usage from historical metrics
   • BExtracting indicators of compromise (IOCs) from a threat intelligence report
   • CClustering similar malware binaries
   • DCalculating the average packet size in network traffic
   Reveal Solution Discussion Select an answer first
4. 43foundation · easy

   What is the primary application of natural language processing (NLP) in cybersecurity?

   • AAnalyzing unstructured text data such as logs and threat intelligence reports
   • BDetecting anomalies in numerical sensor data
   • CGrouping similar network flows
   • DPredicting continuous values such as network latency
   Reveal Solution Discussion Select an answer first
5. 44application · medium

   A security team is deploying an AI-based intrusion detection system. During testing, they notice that the model performs well on the training data but poorly on new, unseen traffic. The team suspects the model has learned patterns specific to the training environment rather than generalizable behaviors. Which limitation of AI analytics is most relevant here?

   • ABias in the training data
   • BLack of explainability
   • COverfitting to the training data
   • DHigh false positive rate
   Reveal Solution Discussion Select an answer first
6. 45expert · hard

   A security analyst is using an AI model that flags anomalous user behavior. The model outputs a risk score, but the analyst does not understand why certain users are flagged. The analyst needs to explain the model's decisions to management and also reduce the number of false positives. Which approach should the analyst take?

   • AUse a model-agnostic explanation technique like SHAP to understand feature contributions, then adjust the threshold
   • BSwitch to a black-box model with higher accuracy
   • CIgnore the risk score and manually review all users
   • DUse a simpler model that is easier to explain, even if it has lower accuracy
   Reveal Solution Discussion Select an answer first