Machine learning provides a foundation for a new paradigm where the facilities of computing extend to the level of cognitive abilities in the form of decision support systems. In the area of water distribution systems, there is an increased demand in data processing capabilities as smart meters are being installed providing large amounts of data. In this paper, a method for multiple-layer data processing is defined for prioritizing pipe replacements in a water distribution system. The identified patterns provide relevant information for calculating the associated priorities as part of a real-time decision support system. A modular architecture provides insights at different levels and can be extended to form a network of networks. The proposed clustering method is compared to a single clustering of aggregated data in terms of the overall accuracy.
Predescu, A., Negru, C., Mocanu, M., Lupu, C., Candelieri, A. (2019). A multiple-layer clustering method for real-time decision support in a water distribution system. In Business Information Systems Workshops BIS 2018 International Workshops, Berlin, Germany, July 18–20, 2018, Revised Papers (pp.485-497). Springer Verlag [10.1007/978-3-030-04849-5_42].
A multiple-layer clustering method for real-time decision support in a water distribution system
Candelieri A.
2019
Abstract
Machine learning provides a foundation for a new paradigm where the facilities of computing extend to the level of cognitive abilities in the form of decision support systems. In the area of water distribution systems, there is an increased demand in data processing capabilities as smart meters are being installed providing large amounts of data. In this paper, a method for multiple-layer data processing is defined for prioritizing pipe replacements in a water distribution system. The identified patterns provide relevant information for calculating the associated priorities as part of a real-time decision support system. A modular architecture provides insights at different levels and can be extended to form a network of networks. The proposed clustering method is compared to a single clustering of aggregated data in terms of the overall accuracy.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.