Evaluate: hidden labels (A) and injected anomalies (B)
10 min applied10 min research2 curated resources
A: precision@k, recall@k and PR-AUC using the hidden fraud labels. B: inject 20 synthetic anomalies (data synthesis from M02) and measure detection.
Concepts: precision@k; recall@k; PR-AUC; synthetic injection; expert review
Project A: Compare IF, LOF, OC-SVM, centroid distance by precision@100/500 and PR-AUC (baseline PR-AUC = 0.0017).
Project B: Which methods catch the injected outage day and the doubled-load hour?
Notes, key code, what to skip and 2 curated resources are inside the platform.Join the waitlist and we'll build you a roadmap that includes this topic only if your goal needs it.