Wikipedia Analysis

Similarly to what hap­pe­ned for the Web at lar­ge, a struc­tu­re has emer­ged from the col­la­bo­ra­ti­ve crea­tion of Wikipedia: its arti­cles con­tain hun­dreds of mil­lions of links. In Wikipedia par­lan­ce, the­se inter­nal links are cal­led wiki­links. These con­nec­tions explain the topics being cove­red in arti­cles and pro­vi­de a way to navi­ga­te bet­ween dif­fe­rent sub­jec­ts, con­tex­tua­li­zing the infor­ma­tion, and making addi­tio­nal infor­ma­tion available.

Cristian Consonni, a for­mer Ph.D. can­di­da­te of mine, has deve­lo­ped novel methods to har­ness  the infor­ma­tion con­tai­ned in the link struc­tu­re of Wikipedia. More pro­sai­cal­ly,  we explo­red the link struc­tu­re of Wikipedia with new approaches.

Three main papers have been published:

[RoyalSociety20] Cristian Consonni, David Laniado, and Alberto Montresor. Cyclerank, or the­re and back again: Personalized rele­van­ce sco­res from cyclic paths on direc­ted gra­phs. Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, 476(2241), September 2020. [PDF[Bibtex] .

[WikiWorkshop19] Cristian Consonni, David Laniado, and Alberto Montresor. Discovering topi­cal con­tex­ts from links in Wikipedia. In Proc. of the Wiki Workshop 2019, May 2019. [Bibtex] .

[ICWSM19]   Cristian Consonni, David Laniado, and Alberto Montresor. WikiLinkGraphs: A com­ple­te, lon­gi­tu­di­nal and mul­ti-lan­gua­ge data­set of the Wikipedia link net­works. In Proc. of the 13th International AAAI Conference on Web and Social MediaICWSM’19, pages 598–607, June 2019. [PDF][Bibtex].

Scroll to top