Networked fuzzy belief rule-based system for spatiotemporal monitoring

This paper introduces a spatiotemporal data aggregation scheme using a novel networked fuzzy belief rule-based (NF-BRB) system. The proposed NF-BRB system is employed to design a decision support tool for relative water quality assessment in the distribution network. Different nodes of the network are grouped in several strata with the connectivity of the nodes shown using a spanning tree. For each stratum, a compact NF-BRB system is designed. The proposed system has flexible and tunable parameters that allow the incorporation of both subjective and numerical information. A learning algorithm to find the locally optimum parameters of the NF-BRB system is employed. The rule aggregation is performed based on a cost-effective approach that uses belief rule utility and disapproval among activated rules. Then, pignistic probabilities of fuzzy evaluation grades are found through aggregation among local strata. The case study of spatiotemporal data aggregation based on hourly data of online monitoring locations of a water distribution network is investigated.