Community detection based on the Matthew effect

Influential nodes identification method based on adaptive adjustment of voting ability

Influential nodes identification technology is one of the important topics which has been widely applied to logistics node location, social information dissemination, transportation network carrying, biological virus dissemination, power network anti-destruction, etc. At present, a large number of influential nodes identification methods have been studied, but the algorithms that are simple to execute, have high accuracy and can be better applied to real networks are still the focus of research. Therefore, due to the advantages of simple to execute in voting mechanism, a novel algorithm based on adaptive adjustment of voting ability (AAVA) to identify the influential nodes is presented by considering the local attributes of node and the voting contribution of its neighbor nodes, to solve the problem of low accuracy and discrimination of the existing algorithms. This proposed algorithm uses the similarity between the voting node and the voted node to dynamically adjust its voting ability without setting any parameters, so that a node can contribute different voting abilities to different neighbor nodes. To verify the performance of AAVA algorithm, the running results of 13 algorithms are analyzed and compared on 10 different networks with the SIR model as a reference. The experimental results show that the influential nodes identified by AAVA have high consistency with SIR model in Top-10 nodes and Kendall correlation, and have better infection effect of the network. Therefore, it is proved that AAV algorithm has high accuracy and effectiveness, and can be applied to real complex networks of different types and sizes.

Finding Central Vertices and Community Structure via Extended Density Peaks-Based Clustering

Community detection is a significant research field of social networks, and modularity is a common method to measure the division of communities in social networks. Many classical algorithms obtain community partition by improving the modularity of the whole network. However, there is still a challenge in community division, which is that the traditional modularity optimization is difficult to avoid resolution limits. To a certain extent, the simple pursuit of improving modularity will cause the division to deviate from the real community structure. To overcome these defects, with the help of clustering ideas, we proposed a method to filter community centers by the relative connection coefficient between vertices, and we analyzed the community structure accordingly. We discuss how to define the relative connection coefficient between vertices, how to select the community centers, and how to divide the remaining vertices. Experiments on both real and synthetic networks demonstrated that our algorithm is effective compared with the state-of-the-art methods.

Making sense of tweets using sentiment analysis on closely related topics

Microblogging has taken a considerable upturn in recent years, with the growth of microblogging websites like Twitter people have started to share more of their opinions about various pressing issues on such online social networks. A broader understanding of the domain in question is required to make an informed decision. With this motivation, our study focuses on finding overall sentiments of related topics with reference to a given topic. We propose an architecture that combines sentiment analysis and community detection to get an overall sentiment of related topics. We apply that model on the following topics: shopping, politics, covid19 and electric vehicles to understand emerging trends, issues and its possible marketing, business and political implications.