Scholarlink

<p>The Influence Maximization process involves identifying the most promising portion of a network by selecting the top ’k’ influential nodes. However, this problem is challenging in both directions that is, whether it selects the most influential nodes or calculates the maximum spread throughout the network. To address this problem, we propose a community-based algorithm. Initially, this algorithm detects overlapping communities and creates a candidate set consisting of nodes that belong to multiple communities. We then calculated the inter-community diffusion strength of each node from the candidate set. Based on their inter-community diffusion strength, we select the top ’k’ influential nodes to form an active seed set and then calculate the total influence spread using a “Monte-Carlo" simulation method. We tested our proposed algorithm on four large-scale social networks and compared it with well-known algorithms to evaluate its usefulness and performance.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cbdp-Im: Community-based diffusion propagation for influence maximization algorithm on large social networks

  • 期刊论文

  • 发表日期：

    2025年5月31日

查看全文
  • Mithun Roy,
  • Indrajit Pan

摘要

The Influence Maximization process involves identifying the most promising portion of a network by selecting the top ’k’ influential nodes. However, this problem is challenging in both directions that is, whether it selects the most influential nodes or calculates the maximum spread throughout the network. To address this problem, we propose a community-based algorithm. Initially, this algorithm detects overlapping communities and creates a candidate set consisting of nodes that belong to multiple communities. We then calculated the inter-community diffusion strength of each node from the candidate set. Based on their inter-community diffusion strength, we select the top ’k’ influential nodes to form an active seed set and then calculate the total influence spread using a “Monte-Carlo" simulation method. We tested our proposed algorithm on four large-scale social networks and compared it with well-known algorithms to evaluate its usefulness and performance.