A data scientist is analyzing audio signals to detect a specific pattern that occurs at irregular intervals. The signal is long (hours) and contains significant noise. The goal is to identify the exact timestamps when the pattern occurs. Which signal processing technique is most appropriate for this task?

Use a matched filter, which correlates the signal with a known template of the pattern to detect its occurrences.

Apply a low-pass filter to remove noise and then visually inspect the signal.

Apply a Fourier transform to the entire signal and look for peaks in the frequency domain.

Use a high-pass filter to remove low-frequency noise and then apply a threshold to detect peaks.

A bank wants to reduce false positives in its fraud detection system while maintaining a high detection rate. The current model uses a simple threshold on a risk score. The bank has a limited budget for manual review of flagged transactions. Which approach best balances the trade-off between false positives and detection rate?

Raise the threshold to flag fewer transactions, reducing false positives.

Use a precision-recall curve to select a threshold that maximizes the F1 score, balancing precision and recall.

Lower the threshold to flag more transactions as fraudulent, increasing the detection rate.

Use a random forest model with default settings, which will automatically balance precision and recall.

Use a clustering algorithm to group traffic and flag points far from cluster centers.

Use a rule-based system with fixed rules for detecting attacks.

Use a static anomaly detection model trained only on normal traffic and never update it.

Use a supervised learning model and retrain it periodically with newly labeled attack data.

CompTIADataX

Domain 5Objective 4