Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Topic-aware Most Influential Community Search in Social Networks

Long Teng, Yanhao Wang, Zhe Lin, Fei Yu

TL;DR

This work defines Topic-aware Most Influential Community Search (TAMICS) on uncertain, topic-based social graphs by integrating a $(k,l,\eta)$-core cohesiveness model with a topic-aware Independent Cascade diffusion framework. It proposes an online algorithm with RIS-based influence estimation and a theory-backed bound, and introduces two index structures—TUC-list and TIE-tree—for substantial query acceleration. The index-based approach delivers up to three orders of magnitude speedups with modest overhead, while experiments show TAMICS yields communities with higher topic relevance, cohesiveness, and influence than state-of-the-art topic-aware or influential CS methods. The results underscore TAMICS’s practical impact for targeted influencer discovery and topic-driven event organization in large social networks.

Abstract

Influential community search (ICS) finds a set of densely connected and high-impact vertices from a social network. Although great effort has been devoted to ICS problems, most existing methods do not consider how relevant the influential community found is to specific topics. A few attempts at topic-aware ICS problems cannot capture the stochastic nature of community formation and influence propagation in social networks. To address these issues, we introduce a novel problem of topic-aware most influential community search (TAMICS) to discover a set of vertices such that for a given topic vector q, they induce a $(k, l, η)$-core in an uncertain directed interaction graph and have the highest influence scores under the independent cascade (IC) model. We propose an online algorithm to provide an approximate result for any TAMICS query with bounded errors. Furthermore, we design two index structures and an index-based heuristic algorithm for efficient TAMICS query processing. Finally, we experimentally evaluate the efficacy and efficiency of our proposed approaches on various real-world datasets. The results show that (1) the communities of TAMICS have higher relevance and social influence w.r.t.~the query topics as well as structural cohesiveness than those of several state-of-the-art topic-aware and influential CS methods and (2) the index-based algorithm achieves speed-ups of up to three orders of magnitude over the online algorithm with an affordable overhead for index construction.

Topic-aware Most Influential Community Search in Social Networks

TL;DR

This work defines Topic-aware Most Influential Community Search (TAMICS) on uncertain, topic-based social graphs by integrating a -core cohesiveness model with a topic-aware Independent Cascade diffusion framework. It proposes an online algorithm with RIS-based influence estimation and a theory-backed bound, and introduces two index structures—TUC-list and TIE-tree—for substantial query acceleration. The index-based approach delivers up to three orders of magnitude speedups with modest overhead, while experiments show TAMICS yields communities with higher topic relevance, cohesiveness, and influence than state-of-the-art topic-aware or influential CS methods. The results underscore TAMICS’s practical impact for targeted influencer discovery and topic-driven event organization in large social networks.

Abstract

Influential community search (ICS) finds a set of densely connected and high-impact vertices from a social network. Although great effort has been devoted to ICS problems, most existing methods do not consider how relevant the influential community found is to specific topics. A few attempts at topic-aware ICS problems cannot capture the stochastic nature of community formation and influence propagation in social networks. To address these issues, we introduce a novel problem of topic-aware most influential community search (TAMICS) to discover a set of vertices such that for a given topic vector q, they induce a -core in an uncertain directed interaction graph and have the highest influence scores under the independent cascade (IC) model. We propose an online algorithm to provide an approximate result for any TAMICS query with bounded errors. Furthermore, we design two index structures and an index-based heuristic algorithm for efficient TAMICS query processing. Finally, we experimentally evaluate the efficacy and efficiency of our proposed approaches on various real-world datasets. The results show that (1) the communities of TAMICS have higher relevance and social influence w.r.t.~the query topics as well as structural cohesiveness than those of several state-of-the-art topic-aware and influential CS methods and (2) the index-based algorithm achieves speed-ups of up to three orders of magnitude over the online algorithm with an affordable overhead for index construction.
Paper Structure (31 sections, 2 theorems, 6 equations, 10 figures, 3 tables, 5 algorithms)

This paper contains 31 sections, 2 theorems, 6 equations, 10 figures, 3 tables, 5 algorithms.

Table of Contents

  1. Introduction
  2. Our Results
  3. Paper Organization
  4. Related Work
  5. Influential Community Search (ICS)
  6. Topic-Aware Community Search (TACS)
  7. Topic-Aware Influence Maximization
  8. Problem Formulation
  9. Graph Notation
  10. Influence Model
  11. Problem Statement
  12. Online Algorithm
  13. $(k, l, \eta)$-Core Computation
  14. Influence Score Calculation
  15. Online Algorithm for TAMICS
  16. ...and 16 more sections

Key Result

Lemma 1

Let $\theta = O(\frac{1}{\epsilon^2} \log{\frac{n}{\delta}})$ and $G'_1, \dots, G'_{\theta}$ be a set of subgraphs sampled from $G_{\bm{q}}$ as Definition def-rrs. For each vertex $v \in V_{\bm{q}}$, we have $\widetilde{\mathbb{I}}_{\bm{q}}(v) = \mathbb{I}_{\bm{q}}(v) \pm \epsilon n$ with probabilit

Figures (10)

  • Figure 1: Workflow of our proposed algorithms for TAMICS query processing.
  • Figure 2: Running examples of a social network with two topics "movie" and "music" and its topic-based interaction graph and influential communities for $\bm{q} = (0.5, 0.5)$.
  • Figure 3: Running examples of the online algorithm for TAMICS with $\bm{q} = (0.5, 0.5)$, $k = 1$, $l = 2$, and $\eta = 0.6$.
  • Figure 4: Examples of a supergraph and the $\eta$-thresholds of vertices in the social network $\mathcal{G}$ of Fig. \ref{['fig_1example:left']} when $k = 1$ and $l = 2$.
  • Figure 5: Illustration of the TUC-list and TIE-tree.
  • ...and 5 more figures

Theorems & Definitions (16)

  • Definition 1: Social Network
  • Definition 2: Topic-based Interaction Graph
  • Definition 3: $(k, l, \eta)$-Core
  • Example 1
  • Definition 4: $(k, l, \eta)$-Influential Community
  • Definition 5: TAMICS
  • Example 2
  • Definition 6: RR Set
  • Lemma 1
  • proof
  • ...and 6 more