| 摘要 |
PROBLEM TO BE SOLVED: To provide a warning generation system on the trend performance analysis. SOLUTION: In this warning generation system (26) on the trend performance analysis, and this method therefor, a multivariate abnormal state detector (28) normalizes the data obtained from a process, and classifies the data in a multidimensional space defined regarding variables in the process. The normalized data is classified into a normal class or an abnormal class. Next, a warming evaluation apparatus (29) evaluates the data classified into the abnormal class, and determines if the data is a true warning or a false one. In another embodiment, the multivariate abnormal state detector (28) and the warning evaluation device (29) are used in parallel with a trend performance analysis tool (18) to validate the warning generated therefrom. |
