{"title":"复杂网络中的拓扑群落","authors":"Luís F Seoane","doi":"arxiv-2409.02317","DOIUrl":null,"url":null,"abstract":"Most complex systems can be captured by graphs or networks. Networks connect\nnodes (e.g.\\ neurons) through edges (synapses), thus summarizing the system's\nstructure. A popular way of interrogating graphs is community detection, which\nuncovers sets of geometrically related nodes. {\\em Geometric communities}\nconsist of nodes ``closer'' to each other than to others in the graph. Some\nnetwork features do not depend on node proximity -- rather, on them playing\nsimilar roles (e.g.\\ building bridges) even if located far apart. These\nfeatures can thus escape proximity-based analyses. We lack a general framework\nto uncover such features. We introduce {\\em topological communities}, an\nalternative perspective to decomposing graphs. We find clusters that describe a\nnetwork as much as classical communities, yet are missed by current techniques.\nIn our framework, each graph guides our attention to its relevant features,\nwhether geometric or topological. Our analysis complements existing ones, and\ncould be a default method to study networks confronted without prior knowledge.\nClassical community detection has bolstered our understanding of biological,\nneural, or social systems; yet it is only half the story. Topological\ncommunities promise deep insights on a wealth of available data. We illustrate\nthis for the global airport network, human connectomes, and others.","PeriodicalId":501043,"journal":{"name":"arXiv - PHYS - Physics and Society","volume":"5 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Topological communities in complex networks\",\"authors\":\"Luís F Seoane\",\"doi\":\"arxiv-2409.02317\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Most complex systems can be captured by graphs or networks. Networks connect\\nnodes (e.g.\\\\ neurons) through edges (synapses), thus summarizing the system's\\nstructure. A popular way of interrogating graphs is community detection, which\\nuncovers sets of geometrically related nodes. {\\\\em Geometric communities}\\nconsist of nodes ``closer'' to each other than to others in the graph. Some\\nnetwork features do not depend on node proximity -- rather, on them playing\\nsimilar roles (e.g.\\\\ building bridges) even if located far apart. These\\nfeatures can thus escape proximity-based analyses. We lack a general framework\\nto uncover such features. We introduce {\\\\em topological communities}, an\\nalternative perspective to decomposing graphs. We find clusters that describe a\\nnetwork as much as classical communities, yet are missed by current techniques.\\nIn our framework, each graph guides our attention to its relevant features,\\nwhether geometric or topological. Our analysis complements existing ones, and\\ncould be a default method to study networks confronted without prior knowledge.\\nClassical community detection has bolstered our understanding of biological,\\nneural, or social systems; yet it is only half the story. Topological\\ncommunities promise deep insights on a wealth of available data. We illustrate\\nthis for the global airport network, human connectomes, and others.\",\"PeriodicalId\":501043,\"journal\":{\"name\":\"arXiv - PHYS - Physics and Society\",\"volume\":\"5 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-03\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Physics and Society\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.02317\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Physics and Society","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.02317","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Most complex systems can be captured by graphs or networks. Networks connect
nodes (e.g.\ neurons) through edges (synapses), thus summarizing the system's
structure. A popular way of interrogating graphs is community detection, which
uncovers sets of geometrically related nodes. {\em Geometric communities}
consist of nodes ``closer'' to each other than to others in the graph. Some
network features do not depend on node proximity -- rather, on them playing
similar roles (e.g.\ building bridges) even if located far apart. These
features can thus escape proximity-based analyses. We lack a general framework
to uncover such features. We introduce {\em topological communities}, an
alternative perspective to decomposing graphs. We find clusters that describe a
network as much as classical communities, yet are missed by current techniques.
In our framework, each graph guides our attention to its relevant features,
whether geometric or topological. Our analysis complements existing ones, and
could be a default method to study networks confronted without prior knowledge.
Classical community detection has bolstered our understanding of biological,
neural, or social systems; yet it is only half the story. Topological
communities promise deep insights on a wealth of available data. We illustrate
this for the global airport network, human connectomes, and others.