This control technique is considered to be very effective in the field of surveillance.
The PCA or PCANL for non-linear parameters is considered as a tool for avoiding or reducing any anomalies that may occur in the field of system reliability.
Dear Lotfi,
Yes we can do it, here are attached files and publications that prove that this possible.
Best regards
Article In-Network PCA and anomaly detection
Article Robust traffic anomaly detection with principal component pursuit
Book Combining soft computing and statistical methods in data ana...
Article A Hybrid Technique Using PCA and Wavelets in Network Traffic...
Conference Paper A semi-supervised model for network traffic anomaly detection
Article PCA-based network Traffic anomaly detection
Article Traffic Anomaly Detection Based on Robust Principal Componen...
Dear Lotfi Nabli,
yes PCA and its descendants (DPCA, KPCA, ...) are very efficient in monitoring large multivariate data.
However, it is recommended to stationary data closer to normal (Gaussian) distribution, and more extensions are proposed to ensure better performance:
https://www.researchgate.net/publication/312511358_A_new_Adaptive_PCA_based_thresholding_scheme_for_Fault_Detection_in_complex_systems
https://www.researchgate.net/publication/317958344_Fault_detection_and_diagnosis_in_a_cement_rotary_kiln_using_PCA_with_EWMA-based_adaptive_threshold_monitoring_scheme
https://www.researchgate.net/publication/319987790_A_modified_Kullback_divergence_for_direct_fault_detection_in_large_scale_systems
Article A new Adaptive PCA based thresholding scheme for Fault Detec...
Article Fault detection and diagnosis in a cement rotary kiln using ...
Article A modified Kullback divergence for direct fault detection in...
- Badges
- Science topic
- Similar topics
- Indexes