This work compares OmniAnomaly with PCA for MTSAD, finding PCA can match or outperform OmniAnomaly.
problem Comparing deep learning models with classical methods in MTSAD under fair evaluation protocols.
method Systematic comparison of OmniAnomaly and PCA on SMD, using identical thresholding and evaluation procedures.
result PCA can achieve performance comparable to OmniAnomaly and even outperform it under certain conditions.