The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
Synthetic Conductance Network Topologies
192 original synthetic undirected connected networks, generated for controlled limited-probe network-repair research. No real infrastructure or personal data. Each network has 24 vertices, 42 edges, and a Hamiltonian cycle before local chords. Edges and vertices are randomly relabeled. Nominal conductance, wear and material covariates are synthetic. topologies.jsonl has one JSON record per graph: graph_id:string (32 hex); partition:string (train/development/test); topology_seed:int; graph:object. graph fields: n_nodes:int; edges:list[[u:int,v:int]] length42 with zero-based vertices; nominal_g:list[float] length42 in [0.6,1.6]; wear:list[float] length42 in [0,1]; material:list[int] length42 in {0,1,2}; ground:int; probe_sources:list[int] length3; sensors:list[int] length4; xy:list[[x:float,y:float]] length24. Partition sizes: 96 train,32 development,64 test. Partitions apply to complete topologies. The release contains no incident observations, failed-edge keys or repair-utility labels. Topology generation seeds reproduce the topology covariates only. A downstream benchmark generates independent incident draws with a retained organizer-private key. Neither benchmark answers nor that key are in this public source. All files are UTF-8. Original synthetic data by Alperen Gur, September 29 2026. CC BY 4.0. This corpus is not calibrated to any real infrastructure system and does not validate maintenance decisions.
- Downloads last month
- 13