Time-delayed cross mapping is a causal feature selection framework that addresses the challenges of time delays and variable interdependence in industrial processes. It uses state space reconstruction to analyze causality, enhancing the accuracy and stability of soft sensor models.
Time-delayed cross mapping is a method for finding cause-and-effect relationships in complex industrial systems, especially when those effects happen over time and variables influence each other. It improves the reliability and accuracy of 'soft sensors' used to monitor these processes by considering these real-world complexities.
TDCM, TDCCM, TDPCM
Was this definition helpful?