[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81521-en":3,"doc-seo-81521-105":29,"detail-sidebar-cat-0-en-105":82},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},81521,1099513958762,"Logic","https://ap-avatar.wpscdn.com/avatar/1000023916a998db790?x-image-process=image/resize,m_fixed,w_180,h_180&k=1784791008015729253",8,"Research & Report","LD-Leiden: Local Parallel Community Detection in Large Dynamic Networks","Dynamic community detection must efficiently update high-quality modularity partitions after edge batches, yet repeating a full Leiden rerun makes tiny changes scale with the entire graph snapshot. Existing dynamic approaches either reduce work but disturb Leiden refinement, store limited hierarchy state, or limit supported graph types. LD-Leiden introduces a local dynamic Leiden method for weighted directed/undirected graphs, updating only repaired affected regions while preserving move-refine-aggregate and enabling parallel local moves with bounded update cost.","LD-Leiden: Local Parallel Community Detection in  \nLarge Dynamic Networks  \nGrigoriy Bokov  \nLomonosov Moscow State University Moscow, Russia  \n1 Leninskiye Gory, 119991 [bokovgrigoriy@gmail.com](bokovgrigoriy@gmail.com)  \nStanislav Moiseev  \nLomonosov Moscow State University Moscow, Russia  \n1 Leninskiye Gory, 119991 [stanislav.moiseev@gmail.com](stanislav.moiseev@gmail.com)  \nAleksandr Konovalov  \nLomonosov Moscow State University Moscow, Russia  \n1 Leninskiye Gory, 119991 [alexandr.konoval@gmail.com](alexandr.konoval@gmail.com)  \nIvan Safonov  \nNational Research University Higher School of Economics Moscow, Russia  \n11 Pokrovksy Bulvar, 109028 [isafonov27@gmail.com](isafonov27@gmail.com)  \nAnna Uporova  \nLomonosov Moscow State University Moscow, Russia  \n1 Leninskiye Gory, 119991 [annuporova2003@gmail.com](annuporova2003@gmail.com)  \nAlexander Radionov  \nMoscow Infocommunication Technology Laboratory Moscow, Russia 1-75b Leninskiye Gory, 119234 [alex.radionov89@gmail.com](alex.radionov89@gmail.com)  \narXiv :2502 . 18497v2 [ cs . SI] 10 Jul 2026  \nAbstract—Dynamic community detection must update highquality modularity partitions after edge batches, yet full Leiden reruns make small changes scale with the whole snapshot. Existing dynamic methods reduce work but often alter Leiden refinement, keep limited hierarchy state, or restrict graph support. This paper presents LD-Leiden, a local dynamic Leiden method for weighted directed and undirected graphs that preserves the move-refine-aggregate pipeline and updates only repaired affected regions. Its novelty is the combination of an affectedfrontier rule after statistic repair, exact subtract-add aggregate repair, and conflict-filtered parallel local moves; together these mechanisms bound update cost by the visited frontier rather than the full graph. On real streams and streamed static graphs with up to 214M vertices and 3.30B edges, LD-Leiden is 48.77× faster than warm-started Leidenalg in 100-batch runs while preserving a 0.996 final modularity ratio. On the common undirected benchmark set, it is 6.94× faster than DF-Leiden and 9.73× faster than NetworKit while obtaining higher final modularity; synthetic sequences support the predicted local edgevolume scaling.  \nIndex Terms—dynamic graphs, community detection, modularity, Leiden algorithm, graph mining, parallel algorithms  \nI. INTRODUCTION  \nLarge networks are rarely static. Social interactions, citation links, transaction graphs, road updates, and communication networks evolve through small batches of edge insertions, deletions, and weight changes. Community detection is a standard tool for summarizing and exploring network structure [1]–[3] . Modularity optimization provides a common objective for this task through its null-model interpretation, standard implementations, and established large-scale benchmarks [4], [5] . Louvain made greedy modularity practical on large graphs [6]; Leiden improved the optimization path by refining communities before aggregation, thereby avoiding poorly connected communities [7] .  \nCorresponding author: Grigoriy Bokov ([bokovgrigoriy@gmail.com](bokovgrigoriy@gmail.com)).  \nThe dynamic version of this problem adds a statemaintenance requirement. A full Leiden rerun after every update provides a conservative quality reference, but it repeatedly scans large parts of the graph even when the batch touches only a small region. Existing dynamic methods reuse the previous partition, screen unaffected vertices, or restrict updates to a dynamic frontier [8]–[12] . These mechanisms reduce repeated work, yet the methods that can be run under a common protocol either remain Louvain-like, drop support for directed graphs, or do not maintain a Leiden refinement hierarchy incrementally. The resulting algorithmic problem is to preserve the refinement and aggregation state used by Leiden while avoiding a full-snapshot scan after every batch.  \nTo address this problem, LD-Leiden is designed as a local parall","cbCaibxEJzyqZeX3","https://ap.wps.com/l/cbCaibxEJzyqZeX3","pdf",11169403,1,11,"English","en",105,"# Introduction\n## Problem: dynamic community detection and Leiden reruns\n## Goal and algorithm design of LD-Leiden","[{\"question\":\"How does LD-Leiden achieve faster updates than warm-started or other dynamic methods?\",\"answer\":\"It combines an affected-frontier update rule after statistic repair, exact subtract-add aggregate repairs using an incremental hierarchy, and conflict-filtered parallel local moves, bounding update cost by the visited frontier rather than the full graph.\"}]",1784173973,28,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":77,"head_meta":79,"extra_data":81,"updated_unix":27},"ld-leiden-local-parallel-community-detection-in-large-dynamic-networks","",{"@graph":35,"@context":76},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/ld-leiden-local-parallel-community-detection-in-large-dynamic-networks/81521/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70],{"name":71,"@type":72,"acceptedAnswer":73},"How does LD-Leiden achieve faster updates than warm-started or other dynamic methods?","Question",{"text":74,"@type":75},"It combines an affected-frontier update rule after statistic repair, exact subtract-add aggregate repairs using an incremental hierarchy, and conflict-filtered parallel local moves, bounding update cost by the visited frontier rather than the full graph.","Answer","https://schema.org",{"og:url":51,"og:type":78,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":80,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":83},[84,88,92,96,101,106,111,114,119,122,126],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":45,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":45,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":112,"slug":113},30,"research-report",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},9,"Religion & Spirituality",20,"religion-spirituality",{"id":117,"doc_module":4,"doc_module_name":45,"category_name":120,"show_sort_weight":117,"slug":121},"World Cup","world-cup",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":123,"slug":125},10,"Lifestyle","lifestyle",{"id":127,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":97,"slug":129},19,"General","general"]