返回
A genetic algorithm using K-path initialization for community detection in complex networks
DOI:10.1007/s10586-016-0698-y.png)
摘要
En 中文
Genetic algorithms have been used in community detection due to their efficiency and accuracy in automatic discovery of communities in complex networks. The traditional method of population initialization for genetic algorithms does not fully consider topology of a network. Therefore, the quality of the initial population may be poor, which may consequently slow down the convergence of the entire process. According to the characteristics of social networks, we propose a K-path initialization method which makes full use of the topological information of a given network. The main focus of our study is to find whether a K-path initialized generic algorithm can bring significant increase in Q value after the FIRST iteration, and to prove whether such algorithm can accelerate the convergence of the entire process for faster community detection over a randomly initialized genetic algorithm. By applying this new algorithm to Karate, Football, and Jazz, we found that the K-path ini-tialized algorithm can increase the Q value by 50-160% on average over that without using K-path initialization after the first round of iteration. This initial advantage is then accumulated during the subsequent iterations by reducing the number of total iterations by at least 28% and at most 41%.
Keyword:
Social networks
Community detection
K-path initialization
Genetic algorithms
AI总结
对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
C
IF:
4.1
论文数:
5.0K
被引数:
7.5K
机构
引用论文
Hydrological soil behavior in areas with semi-arid vegetation (Beni Chougrane Mountains, Algeria)
Biologia
IF0
A multi-task neural network for multilingual sentiment classification and language detection on Twitter用于Twitter多语言情感分类和语言检测的多任务神经网络
Community structure mining in big data social media networks with MapReduce基于MapReduce的大数据社交媒体网络社区结构挖掘

