A Real-time Event-Based Learning Approach for Input-Output Sensitivity Analysis

This paper introduces a real-time event-based learning approach for input-output sensitivity analysis. Based on the rank order clustering method, this proposed approach is capable of continuously monitoring real-time variations in input variables and automatically quantifying each input’s contribution to the target output. Moreover, the method facilitates ranking and grouping based on the significance and correlation of input-output event data. The application to hydraulic systems shows that the developed method not only produces effective evaluation of each input variable’s influence on the target output, but also opens a new avenue for comprehending the performance of hydraulic systems.