You work on an operations team at an international company that manages a large fleet of on- premises servers located in few data centers around the world. Your team collects monitoring data from the

Question

Options

• ATrain a time-series model to predict the machines' performance values. Configure an alert if a
• BImplement a simple heuristic (e.g., based on z-score) to label the machines' historical
• CDevelop a simple heuristic (e.g., based on z-score) to label the machines' historical performance
• DHire a team of qualified analysts to review and label the machines' historical performance data.

• A
  17% (6)
• B
  3% (1)
• C
  74% (26)
• D
  6% (2)

Explanation

Because incident data is not yet labeled, a supervised ML model cannot be trained directly - you first need a labeled dataset. The correct first step is to develop a heuristic (e.g., flag readings that exceed a z-score threshold as anomalous) to programmatically label the historical monitoring data. Those synthetic labels then become the training signal for a supervised model. Option D (hiring analysts) is expensive and slow. Option A (training a time-series model without labels) is not feasible for supervised anomaly detection. Option B describes implementing the heuristic without the subsequent step of using it to train an ML model, which falls short of the stated goal of a predictive ML solution.

No community discussion yet for this question.