Jirí Setinský, Martin Zádník: An Empirical Study of a PCA-Based Multivariate Framework for Interpretable Log Anomaly Detection. CNSM 2025: 1-6