2.1 Recommend AI-powered Data Analytic Techniques to Meet Specific Needs or Answer Specific Questions 350-201 Practice Questions (Page 8)

Questions 36–40

1. 36application · medium

   A security analyst is using a machine learning model that flags suspicious user behavior. The model outputs a risk score between 0 and 1, and the analyst needs to decide which alerts to investigate. The model has high recall but low precision, causing many false positives. The analyst wants to reduce the number of false positives while still catching most true threats. What should the analyst do?

   • AIgnore the risk score and investigate all alerts
   • BRaise the threshold for what is considered suspicious
   • CLower the threshold for what is considered suspicious
   • DUse the risk score as a categorical label without thresholding
   Reveal Solution Discussion Select an answer first
2. 37application · medium

   A security team has a large dataset of 10,000 files, of which 500 are known malware samples and 9,500 are benign. They need to build a model that classifies new files as malicious or benign. The team has labeled data and wants to minimize false negatives (missed malware). Which approach is most appropriate?

   • AAnomaly detection treating benign files as normal and malware as outliers
   • BSupervised classification with a focus on recall for the malicious class
   • CUnsupervised clustering to group files by similarity and then label the clusters
   • DSupervised regression to predict a continuous maliciousness score
   Reveal Solution Discussion Select an answer first
4. 38foundation · easy

   Which scenario is best suited for anomaly detection?

   • APredicting the total number of alerts expected next month
   • BClassifying malware into known families
   • CIdentifying a user who logs in from an unusual location at an unusual time
   • DGrouping similar phishing emails together
   Reveal Solution Discussion Select an answer first
5. 39foundation · easy

   A security analyst wants to discover groups of similar user behavior without any prior labels. Which technique is most appropriate?

   • AClassification
   • BTime-series analysis
   • CRegression
   • DClustering
   Reveal Solution Discussion Select an answer first
6. 40application · medium

   A cloud security team wants to forecast the expected CPU and memory usage of their servers for the next 24 hours to proactively scale resources and avoid performance degradation. They have historical telemetry data with timestamps. Which AI technique should they use?

   • AGraph analytics to map dependencies between servers
   • BClassification to label each hour as 'high usage' or 'low usage'
   • CRegression analysis on historical resource usage data to predict future continuous values
   • DClustering servers by their average usage
   Reveal Solution Discussion Select an answer first