Decision support systems represent specific form of control systems that helpdecision-makers to identify and solve problems, complete decision processtasks and make non-trivial decisions. In context of the process industries thedecision support system (DSS) can help plant operators and engineers to dealwith complex tasks like process monitoring, fault detection and diagnosis, dataanalysis or process optimization. The paper describes specific concept of a datadrivendecision support system that leverages the principle of lazy learning,which builds predictive models locally in the nearest neighborhood aroundgiven point of interest. The methodology of memory-based regression,classification, novelty detection and optimization is described along withpossible applications in the process industries.
展开▼
