Link prediction is widely used in various fields to predict the missing or potential future links between node pairs in the network. Knowledge distillation (KD) methods have been introduced for the link prediction task, demonstrating exceptional performance in inference acceleration. However, these methods primarily focus on supervised learning settings that rely on high-quality labels, and they fail to capture the rich structural features inherent in graph data within the teacher model effectively, resulting in suboptimal performance. To address the problem of reducing label dependency, we propose a link prediction method via contrastive learning with relational distillation (LCLRD) that aims to improve model performance under label scarcity. LCLRD incorporates contrastive learning into the teacher model for pretraining, enabling the teacher model to obtain high-quality node representations more effectively. Then, graph structural information is extracted from the teacher model via relational distillation and transferred to the student model. In addition, we leverage the Pearson correlation coefficient (PCC) as a new matching strategy to replace the Kullback–Leibler (KL) divergence, aiming to alleviate the prediction discrepancy between the student and the stronger teacher model. The experimental results show that LCLRD significantly outperforms other baseline methods on 12 datasets, demonstrating the superiority of this approach.

The diversity and complexity of the user population on the campus network increase the risk of computer virus infection during terminal information interactions. Therefore, it is crucial to explore how computer viruses propagate between terminals in such a network. In this study, we establish a novel computer virus spreading model based on the characteristics of the basic network structure and a classical epidemic-spreading dynamics model, adapted to real-world university scenarios. The proposed model contains six groups: susceptible, unisolated latent, isolated latent, infection, recovery, and crash. We analyze the proposed model’s basic reproduction number and disease-free equilibrium point. Using real-world university terminal computer virus propagation data, a basic computer virus infection rate, a basic computer virus removal rate, and a security protection strategy deployment rate are proposed to define the conversion probability of each group and perceive each group’s variation tendency. Furthermore, we analyze the spreading trend of computer viruses in the campus network in terms of the proposed computer virus spreading model. We propose specific measures to suppress the spread of computer viruses in terminals, ensuring the safe and stable operation of the campus network terminals to the greatest extent.

The diversity structure of the information diffusion and network topology on online social networks were investigated in this article. To accurately capture these phenomena, we combined scale-free and small-world networks to construct a new hybrid network, and the characteristics and evolution law of the network topology and information propagation were probed. The proposed network model is as close as possible to a real network. Further, we constructed hybrid propagation models with susceptible-infected-susceptible, susceptible-infected-recovered, and susceptible-infected-recovered-susceptible models mixed in arbitrary proportions. The resulting model is then used to capture the explosive effect, which refers to the explosive characteristic of information propagation by triggering implicit edges. A threshold analysis, mathematical derivation, and simulation of the new model for simulating hybrid networks were conducted. The results were compared with the power-law distribution of five real network data, demonstrating that our network model can effectively simulate the network generation process. Finally, a simulation experiment on the propagation process was performed, and the results were compared with three sets of real data, confirming that the proposed propagation model can accurately predict the spread of information.

Online social networks serve as an essential platform for human behavior, making the exploration of their impact on epidemic spread a significant issue. This study utilizes the research methodology of propagation dynamics to quantify the interaction between information dissemination in social networks and virus diffusion in physical networks. The double-layer dynamic model was constructed, introducing the concept of dynamic transmission probabilities. The dynamic infection probability of disease transmission incorporates the information immunization effect, where wide-spread information dissemination leads to spontaneous immune behavior among the population. The dynamic transmission probability of information spread incorporates the fear effect, where a surge in mortality rates triggers exponential information dissemination. Mathematical derivations, parameter analyses, and simulation experiments were initially conducted. Subsequently, the differential equations of the double-layer network dynamic model were used to propose the new double-layer dynamic module. Thus, this article proposes a novel epidemic forecasting framework called physics-informed graph neural network (PIGNN). Extensive experiments show that our proposed PIGNN model outperforms the state-of-the-art method. Meanwhile, the double-layer network dynamic model was demonstrated to capture real-world phenomena vividly.

This article explores the role and influence of opinion leaders in the spread of public opinion within social networks and proposes an effective strategy for controlling public opinion accordingly. First, by constructing a cluster social network model based on followee and follower relationships, the structure of real social networks is accurately reflected. Further, a key opinion leader (KOL) indicator system model is proposed to quantify the key opinion leaders’ metrics, and an information forwarding model that collaborates with the KOL indicator system and public opinion orientation is established to simulate information spread in real social networks. The results show that the information forwarding model performs excellently in predicting the spread of public opinion, and the cluster social network model aligns well with the topological structure of real social networks. Without considering positive and negative interventions, the optimal intervention period is when the total spread time proportion is$15$%$\boldsymbol{\leq}\tau\boldsymbol{\leq}32$%. During this period, enhancing or suppressing information spread can achieve the best results. Additionally, it was discovered during the experiments that higher activity levels among opinion leaders might be more important than greater influence for information spread, often affecting the entire public opinion spread process.

There are hierarchical characteristics in real society, and people are often divided according to social attributes, such as salary and power. In order to better explore the laws of information dissemination in the network of hierarchical characteristics, this article first abstractly defines three ways of dissemination of information in the network of hierarchical structure. Then, the information dissemination model of the hierarchical network is constructed, and the model reflects the characteristics that the low-class people are willing to receive news from the high-class people, while the high-class people are less willing to the news from the low-class people. Finally, both the simulated data and the real data have concluded that the middle class is the most active in the uniformly distributed hierarchical network; while the lower class is more active in the power-law distributed hierarchical network, which promotes the dissemination of news in the hierarchical network.

This article explores information spreading in modern online social networks. According to the law of information spread in real social networks, information retweeting is divided into two types: topic retweeting and relationship retweeting. Social behavior is considered as a higher order interaction, and the nodal influence effect in the traditional approach is abstracted as part of it for analysis. The process of topic communities being subjected to social behavior is simulated by the social behavior model, and the dynamic retweeting rate is established. A network evolution model is constructed based on the centrality and noncontinuity characteristics of topic communities in the spread process. The social reinforcement effect in information spreading is described in two dimensions by defining topic expansion rate and topic diffusion rate. This work conducts multiple views of analysis and visualization, which provide more results of quantitative aspect. The validity of the model is verified by comparing the model simulation results with real cases and the generalization ability experiments.

Modern social media networks have become an important platform for information competition among countries, regions, companies, and other parties. This article utilizes the research method of spread dynamics to investigate the influence of the control role of announcements in social networks on the spreading process. This article distinguishes two spreading phases using the authentication intervention as a boundary: the unconfirmed spreading phase and the confirmed spreading phase. Based on the actual rules of spreading in online social networks, two kinds of verification results are defined: true information and false information. The two-stage information spreading dynamics model is developed to analyze the changes in spreading effects due to different validation results. The impact of the intervention time on the overall spread process is analyzed by combining important control factors such as response cost and time sensitivity. The validity of the model is verified by comparing the model simulation results with real cases and the adaptive capacity experiments. This work is analyzed and visualized from multiple perspectives, providing more quantitative results. The research content will provide a scientific basis for the intervention behavior of information management control by relevant departments or authorities.

This article explores the formation and change of group emotion. The phase of group emotion is defined, and the phase transition of group emotion is studied from the following two aspects: the group with and without network structure (the group with coevolution between emotion and network structure). For the group without network structure, the threshold of group emotional phase transition is obtained, and the phenomenon of group emotional polarization is verified. For the group with co-evolution of emotion and network structure, node attractiveness is defined from three aspects: node emotional propensity, node aggregation degree, and node importance. The new network evolution model based on “node attractiveness” is constructed, and the degree distribution of the network is analyzed. The processes and conditions of phase transition in group emotion are obtained based on this evolutionary network analysis. The results show that there are three phases of group emotion: disorder phase, neutral phase, and extreme phase.

As far as a specific rule is concerned, attribute reduction is equivalent to attribute value reduction. Especially for variable precision rough set (VPRS) , any attribute value may be a value reduct, every of them must be checked. Some important information would be ignored if we still first compute attribute reduct and then compute its value reduct just as traditional Rough set theory (RST) in VPRS. This paper, through analyzing the nature of reduction, and the specific meanings of value reduction in VPRS, provide a direct value reduction algorithm.