[Paper Review] Identifying influential spreaders in complex networks

Maksim Kitsak, Lazaros K. Gallos+5 more|arXiv (Cornell University)|Jan 28, 2010

Complex Network Analysis Techniques66 citations

TL;DR

This paper identifies that the most influential spreaders in complex networks are not the most connected or central individuals, but those in the network's core, as revealed by k-shell decomposition. It further shows that inter-spreaders' distance and the persistence of infections in high k-shell layers are critical for efficient information or disease dissemination.

ABSTRACT

Networks portray a multitude of interactions through which people meet, ideas are spread, and infectious diseases propagate within a society. Identifying the most efficient in a network is an important step to optimize the use of available resources and ensure the more efficient spread of information. Here we show that, in contrast to common belief, the most influential spreaders in a social network do not correspond to the best connected people or to the most central people (high betweenness centrality). Instead, we find: (i) The most efficient spreaders are those located within the core of the network as identified by the k-shell decomposition analysis. (ii) When multiple spreaders are considered simultaneously, the distance between them becomes the crucial parameter that determines the extend of the spreading. Furthermore, we find that-- in the case of infections that do not confer immunity on recovered individuals-- the infection persists in the high k-shell layers of the network under conditions where hubs may not be able to preserve the infection. Our analysis provides a plausible route for an optimal design of efficient dissemination strategies.

Motivation & Objective

• To challenge the common assumption that highly connected or central nodes are the most influential spreaders in networks.
• To identify the true structural determinants of influence in complex networks beyond traditional centrality measures.
• To investigate how multiple spreaders interact and how their spatial distribution affects spreading efficiency.
• To explore the persistence of infections in high k-shell layers, especially in non-immunizing scenarios.

Proposed method

• Applying k-shell decomposition to identify the core structure of complex networks, focusing on nodes in the highest k-shells.
• Using epidemic spreading models to simulate information or disease propagation under different initial spreader configurations.
• Measuring spreading efficiency by analyzing the size and duration of outbreaks triggered by different spreader sets.
• Evaluating the impact of inter-spreaders' distance on the overall spreading reach and speed.
• Comparing the performance of high k-shell nodes against nodes with high degree or betweenness centrality.
• Analyzing infection persistence in high k-shell layers under non-immunizing conditions (e.g., SIS model).

Experimental results

Research questions

• RQ1 Do the most influential spreaders in a network correspond to the highest-degree or most betweenness-central nodes?
• RQ2 How does the k-shell decomposition of a network identify more effective spreaders than traditional centrality measures?
• RQ3 What role does the spatial distance between multiple spreaders play in maximizing the spread of information or infection?
• RQ4 Under what conditions can infections persist in high k-shell layers even when hubs fail to sustain transmission?
• RQ5 How does the network core structure contribute to sustained and efficient spreading dynamics?

Key findings

• The most influential spreaders are located in the core of the network, as identified by k-shell decomposition, not by degree or betweenness centrality.
• When multiple spreaders are used, the distance between them becomes the dominant factor in determining the extent of spreading.
• Infections can persist in high k-shell layers even when hubs fail to maintain transmission, especially in non-immunizing scenarios.
• High k-shell nodes form a resilient backbone for spreading, enabling sustained propagation despite the absence of immunity.
• Traditional centrality measures overestimate the influence of hubs, while k-shell core nodes provide a more accurate predictor of spreading efficiency.
• The core structure of a network, revealed by k-shell decomposition, is a superior indicator for designing optimal dissemination strategies.