Possibilistic Bayes modelling for predictive analytics

Author

Carlsson, Christer ; Heikkila, Markku ; Mezei, Jozsef

Author_Institution

IAMSR, Abo Akademi Univ., Turku, Finland

fYear

2014

fDate

19-21 Nov. 2014

Firstpage

15

Lastpage

20

Abstract

Studies in the process industry (and also common sense) show that the most cost effective way to keep production processes running is through predictive maintenance, i.e. to carry out optimal maintenance actions just in time before a process fails. Modern processes are highly auto-mated; data is collected with sensor technology that forms a big data context and offers challenges to identify coming failures from very large sets of data. Modern analytics develops algorithms that are fast and effective enough to create possibilities for optimal JIT (Just-in Time) maintenance decisions.

Keywords

Bayes methods; Big Data; just-in-time; maintenance engineering; big data context; just-in time maintenance decisions; optimal JIT; optimal maintenance actions; possibilistic Bayes modelling; predictive analytics; predictive maintenance; process industry; production processes; sensor technology; Analytical models; Big data; Computational modeling; Data models; Industries; Maintenance engineering; Predictive models;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence and Informatics (CINTI), 2014 IEEE 15th International Symposium on

Conference_Location

Budapest

Type

conf

DOI

10.1109/CINTI.2014.7028671

Filename

7028671