Conflicting evidence combination from the perspective of networks

Combining time-series evidence: A complex network model based on a visibility graph and belief entropy

The random walk-based gravity model to identify influential nodes in complex networks

Link prediction algorithm based on the initial information contribution of nodes

Identification of influential spreaders is still a challenging issue in network science. Therefore, it attracts increasing attention from both computer science and physical societies, and many algorithms to identify influential spreaders have been proposed so far. Degree centrality, as the most widely used neighborhood-based centrality, was introduced into the network world to evaluate the spreading ability of nodes. However, degree centrality always assigns too many nodes with the same value, so it leads to the problem of resolution limitation in distinguishing the real influences of these nodes, which further affects the ranking efficiency of the algorithm. The k-shell decomposition method also faces the same problem. In order to solve the resolution limit problem, we propose a high-resolution index combining both degree centrality and the k-shell decomposition method. Furthermore, based on the proposed index and the well-known gravity law, we propose an improved gravity model to measure the importance of nodes in propagation dynamics. Experiments on ten real networks show that our model outperforms most of the state-of-the-art methods. It has a better performance in terms of ranking performance as measured by the Kendall’s rank correlation, and in terms of ranking efficiency as measured by the monotonicity value.

Identifying influential spreaders in complex networks by an improved gravity model.

Identifying influential nodes in complex networks: Effective distance gravity model

The identification of important nodes in complex networks is an area of exciting growth due to its applications across various disciplines like disease controlling, community finding, data mining, network system controlling, just to name a few. Many measures have thus been proposed to date, and these measures are either based on the locality of nodes or the global nature of the network. These measures typically use distance based on the concept of traditional Euclidean Distance, which only focus on the local static geographic distance between nodes but ignore the interaction between nodes in real-world networks. However, a variety of factors should be considered for the purpose of identifying influential nodes, such as degree, edge, direction and weight. Some methods based on evidence theory have also been proposed. In this paper, we have proposed an original and novel gravity model with effective distance for identifying influential nodes based on information fusion and multi-level processing. Our method is able to comprehensively consider the global and local information of the complex network, and also utilize the effective distance to replace the Euclidean Distance. This allows us to fully consider the complex topological structure of the real network, as well as the dynamic interaction information between nodes. In order to validate the effectiveness of our proposed method, we have utilized the susceptible infected (SI) model to carry out a variety of simulations on eight different real-world networks using six existing well-known methods. The experimental results indicate the reasonableness and effectiveness of our proposed method.

Qiuyan Shang

Papers

Identifying influential nodes in complex networks: Effective distance gravity model

Identifying influential nodes in complex networks: Effective distance gravity model