Liquor-HGNN: A heterogeneous graph neural network for leakage detection in water distribution networks.
Melanie SchallerMichael SteiningerAndrzej DulnyDaniel SchlörAndreas HothoPublished in: LWDA (2023)
Keyphrases
- neural network
- water distribution networks
- detection algorithm
- graph representation
- detection accuracy
- false positives
- graph structure
- genetic algorithm
- graph model
- detection method
- event detection
- automatic detection
- graph construction
- neural network is trained
- neural network model
- directed acyclic graph
- false alarms
- connected components
- computer vision
- object recognition
- artificial neural networks
- object detection
- random walk
- back propagation
- directed graph
- network model
- face detection
- fault diagnosis
- network architecture
- graph theory
- spanning tree
- multi layer perceptron
- bipartite graph
- heterogeneous networks
- feed forward
- real world