Semantic Characterization of Data Series with Application to Facility Control

A. Lisounkin (Germany)


Data-driven knowledge acquisition, semantic coding, approximate pattern matching.


This paper focuses on the use of the data-driven knowl edge acquisition for a phenomenological modeling of water consumers' behavior. Such a modeling addresses the middle-term and long-term facility operation. The data driven knowledge acquisition procedure includes a con text dependent data preprocessing, profile classification, data semantic characterization, profile coding and ap proximate code comparison. The result of the procedure is a process signature a set of one-/multi- dimensional (string, sequential) patterns with semantic variables. The process signature represents typical with respect to opera tion scenarios process profiles. These signature can be interpreted for the scenario based control and the operator training. This procedure was applied for analysis of water consumers' behavior in middle-sized German cities.

Important Links:

Go Back