Bio：Biao Huang, Professor IEEE Fellow, FCAE, FCIC University of Alberta Canada
Abstract：Modernindustries are awash with large amount of data. Extraction of information andknowledge discovery from data for control system design and system monitoring, fromday by day routine process operating data, is especially interesting butchallenging. There exist numerous challenging issues such as unknown causalrelation, nonlinearity, non-Gaussian distributions, high dimensionality,collinearity, multiple modal operations, outlying points, missing measurementetc that must be considered during the information extraction process. The mostfundamental problem is however the causal relation of data, knowledge of whichis essential for data-based control design and fault diagnosis. Thispresentation will discuss concept of causality and state-of-the-art developmentof causality analysis from data. A Python based toolbox for data-basedcausality analysis will also be introduced.
