Sosyal ağlarda örtüşen toplulukların tespit edilmesi
Yükleniyor...
Dosyalar
Tarih
2016
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
İnönü Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Sosyal medya ve ağ yapılarının artan önemi bu konudaki çalışmaları da arttırmıştır. Sosyal ağlar toplulukların bir araya gelmesiyle oluşan yapılardır. Sosyal ağların en genel özelliği, topluluk yapılarıdır. Gerçek ağ yapılarında bir elemanın birden fazla topluluğa dâhil olma olasılığı vardır ve bu duruma örtüşme (overlapping) denir.Bu çalışmada örtüşen topluluk keşfi problemine iki çözüm önerilmiştir. İlk yönteme göre sosyal ağ bir graf olarak modellenmiştir ve bu graftaki her bir tam bağlı alt graf topluluk olarak kabul edilmiştir. Elde edilen sosyal ağın bitişiklik matrisine Bron-Kerbosch algoritması uygulanmış ve yönsüz graftaki tüm maksimal-klikler bulunmuştur. Ardından bu maksimal-klikler revize edilmiş ve önerilen yöntem eşliğinde kesişen toplulukların keşfi sağlanmıştır. Diğer bir yöntemde ise sosyal ağ yine bir graf olarak modellenmiştir. Grafın Laplace matrisi hesaplanmış ve graf özdeğer ve özvektörlerine göre iki gruba ayrılmıştır. Daha sonra minimum kesen ayrıt işlemleri uygulanarak iki grupta da olma ihtimali olan elemanlar tespit edilmiştir.
The growing importance of social media and networking has increased the efforts in this area. Social networks are structures formed by the communities which are came together. The main common feature of all kind of social networks is community structures. In real network structures, an element is likely to be included in multiple groups and this situation is called as overlapping.In this paper, we have two methods for solving the problem of identifying overlapping groups. According to the first method, social network was modeled as a graph and each fully connected subgraphs in this graph has been accepted as a community. Bron-Kerbosch algorithm has been applied to the adjacency matrix of social network modelled as graph and all maximal cliques in undirected graphes has been found. Then, with the suggested method, these maximal cliques was revised so that overlapping communities could be found. In another method, the social network is modeled as a graph again. The Laplacian matrix of graph is calculated and divided into two groups according to its eigenvalues and eigenvectors. Then, the possibility of elements being in two groups is identified by applying the minimum cutting edges process.
The growing importance of social media and networking has increased the efforts in this area. Social networks are structures formed by the communities which are came together. The main common feature of all kind of social networks is community structures. In real network structures, an element is likely to be included in multiple groups and this situation is called as overlapping.In this paper, we have two methods for solving the problem of identifying overlapping groups. According to the first method, social network was modeled as a graph and each fully connected subgraphs in this graph has been accepted as a community. Bron-Kerbosch algorithm has been applied to the adjacency matrix of social network modelled as graph and all maximal cliques in undirected graphes has been found. Then, with the suggested method, these maximal cliques was revised so that overlapping communities could be found. In another method, the social network is modeled as a graph again. The Laplacian matrix of graph is calculated and divided into two groups according to its eigenvalues and eigenvectors. Then, the possibility of elements being in two groups is identified by applying the minimum cutting edges process.
Açıklama
Anahtar Kelimeler
Bilgisayar mühendisliği bilimleri-bilgisayar ve kontrol
Kaynak
WoS Q Değeri
Scopus Q Değeri
Cilt
Sayı
Künye
Karadeniz, E. (2016). Sosyal ağlarda örtüşen toplulukların tespit edilmesi. İnönü Üniversitesi. Malatya.
