Narrow Results

Identifying influential nodes is of theoretical significance in many domains. Although lots of methods have been proposed to solve this problem, their evaluations are under single-source attack in scale-free networks. Meanwhile, some researches have speculated that the combinations of some methods may achieve more optimal results. In order to evaluate this speculation and design a universal strategy suitable for different types of networks under the consideration of multi-source attacks, this paper proposes an attribute fusion method with two independent strategies to reveal the correlation of existing ranking methods and indicators. One is based on feature union (FU) and the other is based on feature ranking (FR). Two different propagation models in the fields of recommendation system and network immunization are used to simulate the efficiency of our proposed method. Experimental results show that our method can enlarge information spreading and restrain virus propagation in the application of recommendation system and network immunization in different types of networks under the condition of multi-source attacks.

Identifying influential nodes in complex networks continues to be an open and vital issue, which is of great significance to the robustness and vulnerability of networks. In order to accurately identify influential nodes in complex networks and avoid the deviation in the evaluation of node influence by single measure, a novel method based on improved Technology for Order Preference by Similarity to an Ideal Solution (TOPSIS) is proposed to integrate multiple measures and identify influential nodes. Our method takes into account degree centrality (DC), closeness centrality (CC) and betweenness centrality (BC), and uses the information of the decision matrix to objectively assign weight to each measure, and takes the closeness degree from each node to be the ideal solution as the basis for comprehensive evaluation. At last, four experiments based on the Susceptible-Infected (SI) model are carried out, and the superiority of our method can be demonstrated.

The real-world network is heterogeneous, and it is an important and challenging task to effectively identify the influential nodes in complex networks. Identification of influential nodes is widely used in social, biological, transportation, information and other networks with complex structures to help us solve a variety of complex problems. In recent years, the identification of influence nodes has received a lot of attention, and scholars have proposed various methods based on different practical problems. This paper proposes a new method to identify influential nodes, namely Attraction based on Node and Community (ANC). By considering the attraction of nodes to nodes and nodes to community structure, this method quantifies the attraction of a node, and the attraction of a node is used to represent its influence. To illustrate the effectiveness of ANC, we did extensive experiments on six real-world networks and the results show that the ANC algorithm is superior to the representative algorithms in terms of the accuracy and has lower time complexity as well.

Identifying influential nodes is a crucial issue in epidemic spreading, controlling the propagation process of information and viral marketing. Thus, algorithms for exploring vital nodes have aroused more and more concern among researchers. Recently, scholars have proposed various types of algorithms based on different perspectives. However, each of these methods has their own strengths and weaknesses. In this work, we introduce a novel multiple attributes centrality for identifying significant nodes based on the node location and neighbor information attributes. We call our proposed method the MAC. Specifically, we utilize the information of the number of iterations per node to enhance the accuracy of the K-shell algorithm, so that the location attribute can be used to distinguish the important nodes more deeply. And the neighbor information attribute we selected can effectively avoid the overlapping problem of neighbor information propagation caused by large clustering coefficient of networks. Because these two indexes have different emphases, we use entropy method to assign them reasonable weights. In addition, MAC has low time complexity O(n), which makes the algorithm suitable for large-scale networks. In order to objectively assess its performance, we utilize the Susceptible-Infected-Recovered (SIR) model to verify the propagation capability of each node and compare the MAC method with several classic methods in six real-life datasets. Extensive experiments verify the superiority of our algorithm to other comparison algorithms.

Finding influential nodes is of significance to understand and control the spreading capacity of complex systems. This paper aims to find influential nodes of bus networks by a proposed node failure process. Network efficiency and average transfer times are used to measure the performance of bus networks. Six node measures including degree, node strength, line number, betweenness, local triangle centrality (LTC) and a measure considering neighborhood similarity called LSS are introduced to evaluate the importance of nodes. Results show that removing nodes with high betweenness value can effectively decrease the network efficiency, but cannot increase the average transfer times. Furthermore, removing nodes with high values of LTC and LSS considering the neighborhood information can damage the bus networks from the perspectives of both network efficiency and average transfer times.

Reasonably ranking the influence of nodes in social networks is increasingly important not only for theoretical research but also for real applications. A great number of strategies to identify the influence of nodes have been proposed so far, such as semi-local centrality (SL), betweenness centrality and coreness centrality, etc. For the sake of ranking more effectively, a new method of identifying influential nodes is proposed in this paper, which takes into account a node’s influence on its neighbors and the node’s position in the network. The influence on neighbors involves two aspects. One is the influence of the target node on its direct neighbors (h-index), the other is the influence on farther neighbors (semi-local centrality). The location of the node in the network is reflected by the improved k-core score, a modified version of k-core index to make it more applicable to practice. Combining both local and global information of node together makes the proposed method a reasonable and effective strategy to identify the influential nodes. The simulation results compared to other well-known methods on six real-world networks demonstrate the effectiveness of the presented method.

The identification of influential nodes is one of the most significant and challenging research issues in network science. Many centrality indices have been established starting from topological features of networks. In this work, we propose a novel gravity model based on position and neighborhood (GPN), in which the mass of focal and neighbor nodes is redefined by the extended outspreading capability and modified k-shell iteration index, respectively. This new model comprehensively considers the position, local and path information of nodes to identify influential nodes. To test the effectiveness of GPN, a number of simulation experiments on nine real networks have been conducted with the aid of the susceptible–infected–recovered (SIR) model. The results indicate that GPN has better performance than seven popular methods. Furthermore, the proposed method has near linear time cost and thus it is suitable for large-scale networks.

Identifying influential nodes is a basic measure of characterizing the structure and dynamics in complex networks. In this paper, we use network global efficiency by removing edges to propose a new centrality measure for identifying influential nodes in complex networks. Differing from the traditional network global efficiency, the proposed measure is determined by removing edges from networks, not removing nodes. Instead of static structure properties which are exhibited by other traditional centrality measures, such as degree centrality (DC), betweenness centrality (BC) and closeness centrality (CC), we focus on the perspective of dynamical process and global structure in complex networks. Susceptible-infected (SI) model is utilized to evaluate the performance of the proposed method. Experimental results show that the proposed measure is more effective than the other three centrality measures.

How to identify influential nodes in complex networks continues to be an open issue. A number of centrality measures have been presented to address this problem. However, these studies focus only on a centrality measure and each centrality measure has its own shortcomings and limitations. To solve the above problems, in this paper, a novel method is proposed to identify influential nodes based on combining of the existing centrality measures. Because information flow spreads in different ways in different networks, in the specific network, the appropriate centrality measures should be selected to calculate the ranking of nodes. Then, an interval can be generated for the ranking of each node, which includes the upper limit and lower limit obtained from different centrality measures. Next, the final ranking of each node can be determined based on the median of the interval. In order to illustrate the effectiveness of the proposed method, four experiments are conducted to identify vital nodes simulations on four real networks, and the superiority of the method can be demonstrated by the results of comparison experiments.

Identifying the influential nodes in complex networks is a challenging and significant research topic. Though various centrality measures of complex networks have been developed for addressing the problem, they all have some disadvantages and limitations. To make use of the advantages of different centrality measures, one can regard influential node identification as a multi-attribute decision-making problem. In this paper, a dynamic weighted Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is developed. The key idea is to assign the appropriate weight to each attribute dynamically, based on the grey relational analysis method and the Susceptible–Infected–Recovered (SIR) model. The effectiveness of the proposed method is demonstrated by applications to three actual networks, which indicates that our method has better performance than single indicator methods and the original weighted TOPSIS method.