کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
7375795 1480075 2018 16 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Overlapping communities detection based on spectral analysis of line graphs
ترجمه فارسی عنوان
تشخیص همپوشانی جوامع بر اساس تجزیه و تحلیل طیفی نمودارهای خطی
کلمات کلیدی
موضوعات مرتبط
مهندسی و علوم پایه ریاضیات فیزیک ریاضی
چکیده انگلیسی
Community in networks are often overlapping where one vertex belongs to several clusters. Meanwhile, many networks show hierarchical structure such that community is recursively grouped into hierarchical organization. In order to obtain overlapping communities from a global hierarchy of vertices, a new algorithm (named SAoLG) is proposed to build the hierarchical organization along with detecting the overlap of community structure. SAoLG applies the spectral analysis into line graphs to unify the overlap and hierarchical structure of the communities. In order to avoid the limitation of absolute distance such as Euclidean distance, SAoLG employs Angular distance to compute the similarity between vertices. Furthermore, we make a micro-improvement partition density to evaluate the quality of community structure and use it to obtain the more reasonable and sensible community numbers. The proposed SAoLG algorithm achieves a balance between overlap and hierarchy by applying spectral analysis to edge community detection. The experimental results on one standard network and six real-world networks show that the SAoLG algorithm achieves higher modularity and reasonable community number values than those generated by Ahn's algorithm, the classical CPM and GN ones.
ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Physica A: Statistical Mechanics and its Applications - Volume 498, 15 May 2018, Pages 50-65
نویسندگان
, , , , ,