Premium
Pattern matching in historical data
Author(s) -
Johannesmeyer Michael C.,
Singhal Ashish,
Seborg Dale E.
Publication year - 2002
Publication title -
aiche journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.958
H-Index - 167
eISSN - 1547-5905
pISSN - 0001-1541
DOI - 10.1002/aic.690480916
Subject(s) - principal component analysis , data mining , similarity (geometry) , matching (statistics) , computer science , multivariate statistics , pattern matching , pattern recognition (psychology) , set (abstract data type) , process (computing) , artificial intelligence , machine learning , mathematics , statistics , image (mathematics) , programming language , operating system
Abstract For many engineering and business problems, it would be very useful to have a general strategy for pattern matching in large databases. For example, the analysis of an abnormal plant condition would benefit if previous occurrences of the abnormal condition could be located in the historical data. A new pattern‐matching strategy is proposed for multivariate time series based on statistical techniques, especially principal‐component analysis (PCA). The new approach is both data‐driven and unsupervised because neither training data nor a process model is required. Given an arbitrary set of multivariate data, the new approach can be used to locate similar patterns in a large historical database. The proposed pattern‐matching strategy is based on two similarity factors: the standard PCA similarity factor and a new similarity factor that characterizes the pattern of alarm violations. An extensive simulation study for a chemical reactor demonstrates that this strategy is more effective than existing PCA methods and can successfully distinguish between 28 different operating conditions.