摘要
三支决策将传统的正域、负域的二支决策语义拓展为正域、负域和边界域的三支决策语义。目前对边界域处理已成为三支决策模型需要解决的一个重要问题,本文在层次粒化社团划分算法框架的基础上,将代价敏感的边界域处理引入社团划分中,给出了一种新的获取非重叠社团划分的方法。基于代价敏感边界域处理的社团发现算法C-TWD在四个典型社交网络数据集karate、football、dolphin和lesmis上取得了优于相关社团划分算法GN、NFA和LPA的性能。同时,本文选取某市移动通讯的实际应用数据,根据通讯基站间是否有信号切换决定两基站节点间是否有边存在来构建网络模型,实验结果表明C-TWD算法同样适用于实际问题的求解。
The three-way decision theory is constructed based on the notions of acceptance, rejection and non-commitment, which can be directly generated by the three regions: positive region POS(X), negative region NEG(X) and boundary region BND(X). In this paper, we introduce cost-sensitive boundary region processing to detect community on basis of hierarchical granulation. Then,Cost-sensitive Three-Way community Detection algorithm Based on boundary region processing(C-TWD) is proposed. Compared with classical detection algorithms(GN, NFA, LPA), the experiments on karate, football, dolphin and lesmis show that our method can gain higher modularity. On other hand,we select partial mobile data to construct actual network. An edge between two nodes means that the signal changes between two nodes. C-TWD also works well on actual network.
出处
《数码设计》
2017年第3期6-9,共4页
Peak Data Science
基金
国家自然科学基金项目(61602003)
关键词
三支决策
层次粒化
代价敏感
边界域处理
the three-way decision
hierarchical granulation
cost-sensitive
boundary regions processing
