As people benefit from the convenience of social networks, they are also exposed to security risks from malicious information. This article proposes a malicious and anti-malicious information propagation dynamics model based on higher-order diffusion networks to address these issues. Firstly, to tackle the problem of individuals' behavior being swayed by the emotional content of malicious information during propagation, we have established an emotional influence mechanism. At the same time, adding user preferences and external drivers proposes a way to calculate individual influence. Secondly, in response to the discrepancies in the propagation efficiency of malicious information among varying users, we establish a higher-order diffusion network and quantify these differences through the stratification of users into three distinct layers of nodes. Additionally, considering the implicit relationship among potential user groups, we establish a group interaction mechanism by mining group attributes and refine the structure of the propagation network. Lastly, given the coexistence and opposition between malicious and anti-malicious information, dynamic game theory is applied to define a state transition equation incorporating anti-malicious information propagators. Consequently, we propose the SAIR model, a novel paradigm for understanding the propagation of malicious and anti-malicious information using higher-order diffusion networks.

As a fundamental task in understanding the information diffusion process, information diffusion prediction has garnered significant attention in recent years. However, most existing studies tend to focus on the structural characteristics of information diffusion while neglecting an important phenomenon: the propagation of topics often carries multiple intentions, which align with specific user cognition at different evolutionary stages. This diversity in propagation intentions and the dynamic nature of user cognition pose challenges for prediction tasks. To address the above issues, this article introduces Buzz, an information diffusion prediction model, by innovatively approaching the problem from the perspective of aligning multiple propagation intentions with dynamic user cognition. First, to tackle the dynamic hierarchy of propagation intentions, a concise and efficient cascade intention extraction module is designed. This module uses observed diffusion cascades as intention anchors and employs an improved self-attention mechanism to generate representations of the current multiple propagation intentions. Based on this, attention weights are utilized to dynamically stratify the hierarchical structure of propagation intentions. Second, to address the dynamic nature of user cognition, we take social relationships as cognition anchors to initialize the latent diffusion network and dynamically weight the adjacency matrix through temporal slicing. This accurately models the dynamic diffusion process of topics. On this foundation, the cascades are segmented along the diffusion timeline into corresponding time slices, and graph convolution is applied to refine user dynamic cognition. Finally, considering the complexity of aligning propagation intentions with user cognition, we design a multihead attention fusion module. This module aligns propagation intentions based on user cognition, enabling more precise selection of target users. The great performance on four public datasets validates the effectiveness of our proposed approach.

The spread of hot topics in social networks exhibits strong contagiousness, reflecting the interplay between emotional resonance and social influence. Understanding this mechanism is crucial for explaining collective online behavior and improving information diffusion management. This study proposes a hot topic propagation model based on short-term emotions and behavioral differences. Traditional sentiment analysis methods rely excessively on single-text data. To address this limitation, the proposed model quantifies the matching degree between user emotions and topic content through nonlinear functions. In addition, it constructs sentiment influence factors from multimodal information to reveal the emotional mechanisms of communication. Considering the uncertainty of user behavior and the driving effect of group influence, an intention-driving model grounded in communication psychology is developed to jointly describe individual and collective behavioral tendencies. Furthermore, a behavioral difference-based propagation model, SCOIR, is established to characterize the temporal dynamics and behavioral complexity of topic diffusion through information entropy and a reward matrix. Experimental results demonstrate that the proposed model effectively captures users’ emotional dynamics and behavioral patterns, achieving accurate simulation of topic propagation. Overall, this work introduces a multimodal sentiment influence modeling approach, an intention-driving mechanism based on communication psychology, and a behaviorally heterogeneous propagation model SCOIR, providing both theoretical insights and practical implications for understanding emotion-driven information diffusion.

In contemporary society, online public opinion exerts significant influence, and guiding its trajectory is essential for maintaining social stability. To capture the interaction and evolution of competing guiding information during topic dissemination, a competitive guidance topic propagation model based on evolutionary game theory is developed. First, to address the interactive influence of multiple topic features, a multifactor analysis framework is established. Based on positive and negative sentiment, a metric for topic controversy is designed. Considering the effectiveness of multiple regression in measuring diverse features, multiple regression is introduced to evaluate the driving forces of different topic characteristics. This approach allows a detailed analysis of how various features affect both the dissemination impact and the dissemination potential. Second, to address the competition and cooperation among different types of guidance information, evolutionary game theory is introduced to design a user participation-driven game mechanism. The mechanism is used to analyze the competitive and cooperative dynamics between diverse guidance messages. It reveals the strategic landscape of positive and negative information dissemination and provides a theoretical foundation for modeling user state transitions. Finally, to address the dynamic nature of user state changes and considering the similarity between infectious disease transmission and topic propagation, a suspicion–hesitation–positive and negative information–immunity model is proposed. The model allows fine-grained modeling of user state transitions and provides effective theoretical support for public opinion management. Experimental results demonstrate that the proposed model effectively fits the propagation trends of guiding topics on the given dataset and captures competitive and cooperative relationships among different types of guiding information. Compared with baseline models, performance improves by at least 2.8%, providing a solution for modeling and analyzing the propagation process of guiding topics.

Guided topic detection plays a crucial role in public opinion management and crisis response. To address the adversarial nature of group stances and the dynamic evolution of topics, this study proposes a guided topic detection model based on topic evolution and group stances. First, considering the hidden nature of groups in guided topics, the TA-Louvain hidden community mining method is proposed to quickly and accurately discover the community structure and track the structural changes of the same group. Then, for the complexity of group features in guided topics, an attention mechanism-based group feature representation method AG2vec was designed, which can improve the accuracy and robustness of group feature representation. Finally, to address the confrontational nature of group stances, stance-related features are extracted by combining game theory with a multiple linear regression algorithm. Based on this, group change GRU (GC-GRU), a GRU-based guided topic detection model, is proposed to capture the dynamic evolution of group features over time. Experimental validation shows that the method can flexibly respond to the rapid evolution of topics. It also fully accounts for the confrontational and dynamic types of user group stances, thereby enabling the effective detection of guided topics.

This article investigates an emerging social commerce model that integrates social networks and e-commerce, highlighting the pivotal role of topic marketing in influencing consumers’ purchasing decisions. Within the framework of user-generated content (UGC), and accounting for the complex interplay among factors influencing topic propagation as well as the cooperative and competitive dynamics between topics, this study proposes a marketing topic propagation model based on factor analysis and the Stackelberg game. First, factor analysis was employed to extract two primary factors influencing topic propagation—user behavior and external environmental factors—thereby reducing the dimensionality of the variables. A multiple linear regression model was subsequently applied to quantify the contribution of these factors, establishing a robust quantitative foundation for optimizing marketing strategies. Second, to account for the competitive and cooperative interactions among topics, the behavior of ordinary users and influencers was analyzed using a Stackelberg game framework. This analysis reveals the complex interdependencies between cooperative and competitive dynamics in topic propagation. Finally, to address the behavioral variability and time-varying nature of users’ active participation or neglect (of topics), the traditional susceptible-infectious-recovered (SIR) model was refined by subdividing the infectious state into “marketing” and “neglected” states, resulting in the susceptible-marketing-neglected-recovered (SMNR) dynamic propagation model. By incorporating a temporal dimension, the model captures dynamic changes in states and parameters during topic propagation, thereby enhancing the accuracy of predictions regarding propagation paths and trends. Experimental results demonstrate that the proposed model effectively measures dynamic changes in influencing factors, incorporates game-theoretic relationships, and accurately predicts the temporal evolution of marketing topics. Its dynamic characteristics provide a robust theoretical foundation for optimizing social commerce marketing strategies and offer precise guidance for data-driven marketing decisions.

Aiming to optimize marketing promotion, an information dissemination model integrating trust and user preference is developed. The objective is to capture users’ behavioral mechanisms and enhance marketing decision-making efficiency. To measure users’ trust in key opinion leaders, an Interval Type-2 Fuzzy Sets (IT2FSs) -based trust evaluation model is created, enabling effective trust assessment and stimulating purchasing behavior. Regarding the dynamic nature of user preferences, a Hidden Markov Model (HMM) -based prediction algorithm is proposed to track interest changes and forecast repurchase behavior. Considering rational and irrational user behaviors in marketing information dissemination, two new states, repurchase state P and hesitant purchasing state H are introduced based on the Susceptible-Infected-Recovered (SIR) model. Then, the SIRPH social platform information dissemination model is constructed, achieving accurate prediction and enhancement of marketing information dissemination. Experimental results indicate that the SIRPH model reduces the peak purchasing users by 15–25%, extends topic lifecycles by 20–30%, and improves information spreading accuracy, demonstrating the effectiveness of trust and preference integration.

Client selection can result in more homogeneous data and improve the generalization performance of the global model. In the actual training environment, the heterogeneity of client devices and the non-IID characteristics of data seriously reduce the performance of the model. Therefore, this paper proposes a federated learning cluster selection model to alleviate client drift. To address the issue of unstable computing and communication resources among participants, we propose an algorithm that adaptively adjusts the task load based on the linear upper confidence bound algorithm (LinUCB). The algorithm mines the probability distribution of resources from historical training data,whilch helps alleviate the problem of dropout caused by device overload. To solve the high computational complexity problem caused by high-dimensional gradients, a method for extracting the updated direction of a local model is proposed based on updated weights. The updated direction is jointly represented by the class distribution calculated from the cross-entropy loss of the model on a balanced dataset, and the updated weights that are reduced in dimensionality. Finally, we propose a client clustering selection algorithm based on the updated direction to address the problem of client drift in collaborative training with heterogeneous data. The algorithm prioritizes selecting clients whose update directions are close to that of the global model, while simultaneously selecting clients with tilted update directions with equal preference. This makes the data distribution more homogeneous and alleviates the problem of client drift. Experimental results demonstrate that the proposed algorithm for adaptive task allocation can reduce the number of dropouts under resource-limited and unstable scenarios. The client selection algorithm shows good performance on three public datasets, MNIST, CIFAR-10 and AG_NEWS, outperforming six baseline methods across varying levels of data heterogeneity.Our code is public athttps://github.com/Wannery/MBUT-CS.

In the context of social commerce, user dissemination behavior is influenced by the coupling of multiple factors, such as the timeliness of marketing content and the dynamics of social relationships. Existing methods are limited in predictive accuracy due to neglecting the cold-start data sparsity and the evolution patterns of behavior sequences. To address this issue, we propose a predictive model for user dissemination behavior in social commerce, leveraging transfer learning and sparse representation techniques. Firstly, to tackle the problem of sparse cold-start topic data, a cross-domain transfer learning framework is designed. Historical topics are matched as the source domain based on semantic similarity, and knowledge transfer is achieved through a domain adaptation loss function, thereby improving the model's performance for cold-start or low-engagement topics. Second, to address the challenge of high-dimensional feature modeling, a dual sparse coding method is proposed. This method performs low-rank decomposition on both the user attribute matrix and the social adjacency matrix, preserving core structural information while reducing computational complexity. Finally, to handle the dynamic propagation characteristics, a utility function combining tripartite game theory is constructed, incorporating the marketing content attractiveness, user social influence, and time decay factor. This is embedded into the weight generation module of the relationship-aware graph attention network. By using a dynamic time encoder, the evolution pattern of behaviors is captured, enabling accurate prediction of social commerce user behavior. Experimental results show that this model can not only accurately predict users' interactions with marketing topics but also reveal the dissemination patterns of marketing content, providing strong support for optimizing marketing strategies. This research offers an effective technical means for the precise placement of marketing content and user behavior analysis in social commerce.

The development of social networks has prompted a shift in marketing strategies, with a surging demand for marketing in vertical domains characterized by high user stickiness and specialization. To address this, we propose a traceability model based on domain preference and heterogeneous networks. Firstly, considering the problem of marketing topic vertical domains features metric and the influence of users' preference degree for domains on topic propagation, the domains are treated as latent semantics, and the user-topic association matrix sparse matrix is densified using a latent factor model to mine the domain preference information efficiently. Secondly, considering the complexity of the association between multi-type elements in marketing topics, the HLN2vec (Heterogeneous Layer- wise Networks) model is proposed. This model uses heterogeneous network representation learning and incorporates multi-layer attention networks to learn the representations to portray a marketing topic's key elements and their relationships. Finally, this paper proposes the DP-Rank(Domain Preference-based) algorithm, which uses domain preference features and an adaptive random walking strategy to quantify element influence. Based on experiments, the proposed model robustly applies in social networks and exhibits clear advantages in measuring vertical domain features of marketing topics, constructing multi-type element relationship networks, and discovering core element influence.

Vertical federated learning can aggregate participant data features. To address the issue of insufficient overlapping data in vertical federated learning, this study presents a generative adversarial network model that allows distributed data augmentation. First, this study proposes a distributed generative adversarial network FeCGAN for multiple participants with insufficient overlapping data, considering the fact that the generative adversarial network can generate simulation samples. This network is suitable for multiple data sources and can augment participants' local data. Second, to address the problem of learning divergence caused by different local distributions of multiple data sources, this study proposes the aggregation algorithm FedKL. It aggregates the feedback of the local discriminator to interact with the generator and learns the local data distribution more accurately. Finally, given the problem of data waste caused by the unavailability of nonoverlapping data, this study proposes a data augmentation method called VFeDA. It uses FeCGAN to generate pseudo features and expands more overlapping data, thereby improving the data use. Experiments showed that the proposed model is suitable for multiple data sources and can generate high-quality data.

Cross-domain recommendation can combine users' interests in multiple domains to achieve diversification of recommended content. To solve the problem of complex and changeable interest relationship among different domains, this paper proposes a cross-domain recommendation model integrating user's long-term and short-term interests. Firstly, to solve the problem of different data distributions in different domains, this paper proposes the graph convolutional network with maximum mean discrepancy (MMD), which gradually aligns the feature representation of shared potential space during training to reduce the influence of negative transfer. Secondly, to solve the problem of different features of items in different domains, considering the advantage of attention mechanism in distinguishing importance, a hierarchical attention mechanism focusing on item features is proposed to build users' short-term interests in source domain. Finally, aiming at the dynamic influence of users' long-term and short-term interest, a multi-type interest dynamic aggregation module based on gated network is proposed, which not only integrates users' preferences in different domains, but also takes into account the characteristics of items to be predicted. Experiments show that the model can effectively capture users' multi-type interests in source domain and target domain on real data set, so as to improve the recommendation accuracy of target domain.

As topics evolve during the dissemination process, their derivative characteristics play an important role in revealing the mechanism of topic dissemination in social networks. Due to users’ cognitive inertia, followers of antecedent topics tend to pay more attention to derivative topics than ordinary users, and this cognitive inertia is directly related to the correlation between these two topics. Based on this finding, we propose a derivation topic dissemination model based on iterative multitopic derivation and social psychology, taking into full consideration users’ emotional accumulation of antecedent topics and repeated derivation of topics. First, a multiple linear regression model is used to construct a metric algorithm for user antecedent sentiment and to effectively analyze the dynamics of antecedent sentiment accumulation affecting the spread of derivative topics. Second, to analyze the interactions between and within multitopic layers in full, an iterative multitopic dissemination model iterative-susceptible infectious recovery (SIR) is proposed. In addition, the association degree is introduced to define the cross-model state transition equation by considering the association and differences among multitopics. Last, considering the influence of cognitive inertia and “continuous attention psychology” in social psychology, we construct a user psychology-based driving force model which can further improve the cross-model state transition equation. According to the experiments, the model can effectively reveal the influence of different factors on the dissemination trend of derivative topics in social networks, as well as depict the dissemination dynamics of derivative topics.

In topic dissemination, accurately predicting the behavior of derived topic dissemination models enables the implementation of timely measures to effectively mitigate negative impacts. Addressing the issues of data scarcity and the influence of leaders on users, this article proposes a guided derived topic propagation model based on topic identity and transfer learning. To tackle the sparsity of early data on derived topics, we introduce generative adversarial networks (GANs) to enhance data integrity and continuity, given their efficacy in compensating for missing data. Additionally, recognizing the potential for domain shifts during topic evolution, we incorporate transfer learning to facilitate domain adaptation and improve the accuracy of the derived topic propagation model. To address the significant impact of leader authority on user perspectives, we incorporate the social psychological concept of “identity theory” and design a metric for topic identity based on this theory to quantify the leader’s influence on user viewpoints. Concurrently, considering user identification with topics, we propose a metric for topic domain co-presence, incorporating interest preferences to more accurately represent user engagement with the topic. Finally, to account for the dynamic nature of user viewpoints, we employ graph convolutional networks (GCNs) to model user perspectives over time. The use of GCNs, based on temporal slices, effectively handles the interconnections between user nodes and explores the influence of interuser interactions on viewpoints. Experimental results demonstrate that the proposed model not only enhances data integrity and continuity but also improves the prediction of the propagation dynamics of guided derived topics.

The propagation of hot topics often gives rise to a series of derivative topics. In view of the sparsity of user behavior data and the cognitive accumulation of the original topic, a prediction model of derived topic propagation based on cognitive accumulation and transfer learning is proposed. First, for the complexity of the derived topic feature space, considering the relation and difference between derivative topics and original topics, this study designs I(Iterative)T(Topic)2vec, a topic iterative representation method based on original topics to get the low-dimensional representation of the derived topic feature space more richly from the perspectives of both original topics and derivative topics. Second, it aims at the problem of users’ cognitive accumulation of the original topic before the outbreak of derivative topic. The subjective game theory is introduced to construct the cognitive influence of users. At the same time, considering the timeliness of the propagation cycle of derivative topics, we discretized the derivative topic data, and further proposed a derivative topic propagation model based on Subjective Adapt-CNN (SA-CNN). Finally, the sparsity of effective behavior data of users at the beginning of the outbreak of derivative topics is discussed. Considering the rich user behavior data in the communication history of the original topic, data migration is carried out by using the original topic. At the same time, the domain adaptive method based on Transfer Component Analysis (TCA) is introduced to achieve feature adaptation from the original topic data to the derived topic data, further improving the accuracy of the derived topic propagation model. Experiments show that this model can not only effectively alleviate the problem of data sparsity but also perceive the propagation situation of derived topics well.

Discovering hidden key users of leading topics plays an important role in opinion control and risk prevention. Aiming at the dynamic nature of key users’ intentions and other problems, a key user discovery model based on behavioral intentions and implicit relationships is proposed. First, to address the dynamic nature of key users’ intentions, the dynamic latent Dirichlet allocation method is introduced. This approach effectively mines topic evolution in text data, uncovering dynamic behavioral themes of key users and analyzing evolutionary relationships between topics. Meanwhile, incremental learning is introduced to quantify the dynamic behavioral intentions of key users more precisely. Second, a random wandering strategy based on user interaction degree and propagation depth is designed to address the hidden nature of user relationships. The strategy introduces the user interaction degree designed by the social cognition theory and the propagation depth designed by the propagation chain theory to better explore the hidden user interaction relationships. Finally, for the timeliness of key user identification, considering the advantage of dynamic evolution for real-time interaction, dynamic evolution is introduced to effectively analyze the dynamic structure of topic networks, and attention mechanism is introduced to improve the adaptivity of the model. The experiments show that this paper verifies the factuality of the existence of hidden key users dominating the promotion behind the guiding public opinion, and is more effective in tracing the hidden key users in the topics.

Federated learning is an important distributed machine learning paradigm. This study proposes a privacy-preserving data augmentation model for federated learning of heterogeneous data, which is able to mitigate heterogeneity and augmenting the participant’s local data while protecting data privacy. First, to address the problem of global model bias due to heterogeneous data, this study proposes a distributed generative adversarial network FedEqGAN. The model introduces a multi-source data feature fusion mechanism, which can learn the features of each data source to generate synthetic data. Second, addressing the privacy leakage issue caused by the disclosure of data distribution information, this paper proposes an encryption algorithm for heterogeneous environments FedHE, which utilizes homomorphic encryption to protect local data distributions and aggregates local data information through KL dispersion in order to construct global data distributions. Finally, for the privacy leakage problem caused by uploading model parameters in federation training, this paper proposes a federation model parameter encryption algorithm DPFedMP. This algorithm dynamically injects Gaussian noise into the model parameters according to the difference of data distribution to realize differential privacy protection and update the global model. Experiments show that the method is applicable to heterogeneous data environment, significantly enhancing model performance while ensuring data security.

The spread of malicious information is highly destructive to society. In this article, we propose a prediction model for malicious information dissemination based on user behavior analysis. We analyze the potential connection between users and malicious information and the implicit relationship between features. First, considering the complex relationship between users and malicious information, and the fact that knowledge learning can effectively capture this dynamic, this article proposes a method to represent the implicit connection between users and information as vectors, in which the Trans-H knowledge representation learning algorithm is used to map the implicit relationships between users and malicious information onto a low-dimensional relational plane, achieving vectorized representation. Second, considering the interactions between different user features, this article uses a behavioral analysis network to solve the problem of insufficient expression ability of users' original features. This network can be used to predict users' preforwarding behaviors. Finally, time slicing is introduced for the timeliness of malicious information, and the diffusion stage of information is discretized, while the behavioral analysis network is populated according to the user's preforwarding behavior, and a kind of information-user feature matrix is constructed. Then, combined with the convolutional neural network on local spatiotemporal features, this article proposes a prediction model TS-CNN. Experiments show that on publicly available benchmark datasets, comparative experiments with existing popular methods reveal that our approach achieves significant improvements in accuracy, recall, and F1-score, thereby validating the effectiveness of our proposed relationship modeling and optimization method. The model cannot only effectively predict users' dissemination behaviors for information but also more realistically reflect the hidden factors that drive users' dissemination.

In social networks, studying methods for predicting the popularity of malicious information can help improve the ability to predict online public opinion. This paper proposes a malicious information popularity prediction model based on user influence, targeting the cooperative adversarial nature of malicious information propagation, the problem of assessing user influence in malicious information propagation space, and the complexity of malicious information propagation space. First, regarding the cooperative adversarial nature of the malicious information propagation process, considering that user behavior is influenced by both malicious and positive information during the propagation process, evolutionary game theory and multiple linear regression are introduced, and internal and external behavioral factors of the user are synthesized to construct influential functions that quantify malicious information and positive information. Meanwhile, the influence matrix is introduced when quantifying information to construct a weighted malicious information propagation network further. Second, regarding the problem of assessing user influence in the malicious information propagation space, considering the advantages of PageRank in measuring the importance of web pages and combining the timeliness of malicious information propagation, an improved algorithm T-PageRank (Timeliness-PageRank) based on timeliness is proposed. Introducing the time decay factor into the PageRank algorithm effectively enhances the accuracy and timeliness of the influence assessment of malicious information propagation. Finally, regarding the complexity of the propagation space of malicious information and considering that Graph Attention Network (GAT) can effectively capture complex relationships between nodes, combined with user influence, a malicious information popularity prediction model based on GAT is constructed. The model learns the complex interaction between users by using GAT and updates the feature representation of users so that it can be used for subsequent malicious information popularity prediction tasks. The experiment shows that the model can not only accurately assess the influence of users but also effectively predict the popularity of malicious information propagation.

The open and free nature of online platforms presents challenges for tracing malicious information. To address this, we propose a traceability model based on neighborhood similarity and multitype interaction. First, we propose neighborhood similarity algorithms (D-NTC) to address the universality of malicious information dissemination. This algorithm evaluates the impact of user node importance on malicious information propagation by combining node degrees and the topological overlap of neighboring nodes. Second, we consider the interactive nature of multiple types of elements in the network and construct an interactive module based on user-path-malicious information. This module effectively captures the mutual influence relationships among diverse elements. Additionally, we employ representation learning to optimize the transition probability matrix between elements, leveraging hidden relationships to further characterize their interactive impact. Finally, we propose the NSMTI-Rank algorithm, which tackles the complexity of quantifying the influence of multiple types of elements. Drawing inspiration from mutual reinforcement effects, NSMTI-Rank comprehensively quantifies element influence through an iterative framework. Experimental results demonstrate the effectiveness of our approach in mining user node importance and capturing the interaction information among diverse elements in the network. Moreover, it enables the timely and effective identification of sources of malicious information dissemination.

This work is inspired by high-definition (HD) image generation techniques. When the user's interests are viewed as different frames of varying clarity, the unclear parts of one interest frame can be clarified by other interest frames. The user's overall HD interest portrait can be viewed as a fusion of multiple interest frames through detail compensation. Based on this inspiration, we propose a model for generating HD interest portrait called interest frame for recommendation (IF4Rec). First, we present a fine-grained pixel-level user interest mining method, Pixel embedding (PE) uses positional coding techniques to mine atomic-level interest pixel matrices in multiple dimensions, such as time, space, and frequency. Then, using an atomic-level interest pixel matrix, we propose Item2Frame to generate several interest frames for a user. The similarity score of each item is calculated to fill the multi-interest pixel clusters, through an improved self-attention mechanism. Finally, stimulated by HD image generation techniques, we initially present an interest frame noise compensation method. By utilizing the multihead attention mechanism, pixel-level optimization and noise complementation are performed between multi-interest frames, and an HD interest portrait is achieved. Experiments show that our model mines users' interests well. On five publicly available datasets, our model outperforms the baselines.

Recommendation systems greatly improve user experience with personalized suggestions. Much research utilizes deep learning to extract user interest features, yet often neglects their real-time and dynamic aspects. This paper proposes a joint learning-based recommendation model (JLS-Rec) for e-commerce platforms, integrating both short-term and long-term user interests. First, for users' long-term interests, the paper suggests mining more detailed interests from user behavior sequences. This method decouples the behavior sequence in horizontal and vertical directions using a Convolutional Neural Network to learn the user's long-term interests. Second, for users' short-term interests, the JLS-Rec method learns feature transformations on users' recent behavior sequences by stacking multiple self-attention mechanisms, resulting in dynamic representations of the user's short-term interests at the current stage. Finally, based on the principle of prioritizing short-term memory without neglecting long-term interests, the paper proposes a joint learning framework with dual embeddings to balance the two characteristics of user long-term and short-term interests. This framework generates accurate recommendation results while utilizing these two interest features to predict user feedback on products. The experimental results demonstrate that the model effectively mines the long-term and short-term interest information of users in the features, thereby improving the recommendation accuracy of e-commerce platforms.

In online social networks, rumors have become a flashpoint for public opinion. This article proposes a rumor and anti-rumor diffusion model based on multi-information multidimensional compound game considering the complex game antagonism in the process of spreading multiple messages. First, in view of the diversity and complexity of the characteristic space of communication individual, communication structure, and multitype rumor and anti-rumor news, the communication network is expressed in low-dimensional, real-value, and dense vectorization by combining the representation learning algorithm and centering on the content, structure, and attribute of information transmission. Second, considering the antagonistic and competitive relationship between multiple rumors and anti-rumors, and considering the compound iterative cascade of multi-information rumor, this article proposes a multi-information rumor and anti-rumor game model from a multidimensional perspective. Finally, the information expression of the topic space and the complex game relationship between multiple messages are considered comprehensively. At the same time, considering the graph convolutional network (GCN)’s ability of convolution processing non-Euclidean data such as social network, a dynamic, unified representation, multimessage complex game topic propagation model MMGameGCN is proposed based on graph convolutional neural network (MultiMessage-game GCN). The experiment shows that the model can more truly and effectively reflect the process of network rumor propagation in the multimessage transmission network, effectively analyze the user group behavior of rumor topic under multiple messages, and correctly perceive the spread situation of rumor and anti-rumor.

This study is inspired by the current image restoration technology. If we regard the users participating in the rumor as image pixels, similar to social networks, the recovery of pixel data is affected by the pixels themselves and neighbor pixels, then the prediction of user behavior in the rumor diffusion can be regarded as the process of image restoration for pixel-blurred user behavior images. We first propose a diffusion2pixel algorithm that transforms the user relationship network of topic diffusion into image pixel matrix. To cope with the diversity and complexity of the diffusion feature space, the user relationship network is reduced to a low-rank dense vectorization by representation learning before being pixelated by cutting and diffusion. Second, considering the competitive relationship between rumor and anti-rumor, transition matrix of rumor mutual influences is established by evolutionary game theory. A mutual influence model of rumor and anti-rumor is then proposed. Finally, we combine the transition matrix of rumor mutual influence into a simple prediction method Graph-CNN of rumor and anti-rumor topic diffusion based on dynamic iteration mechanism. Experiments confirmed the proposed model can effectively predict the group diffusion trends of rumor, and reflects the competitive relationship between rumor and anti-rumor.

With the increasing network service content, innovative methods are required for developing optimized network service for e-commerce companies. Accordingly, this study focuses on designing a framework containing personalization, interest group identification, and recommendation mechanisms. The primary contribution of this paper is to propose a recommendation model based on data compensation and dynamic user interest grouping. First, to address the problem of sparse user rating data, homeostasis compensation is performed on native data to more realistically restore the preference relationship between users and items by introducing the advantages of generative adversarial network in learning data distribution and enhancing data samples. Second, to address the problem of user interest generalization, information entropy is introduced to measure the user interest feature space. In addition, the time window marking method is used to further quantify the users’ dynamic interest group around the users’ interest drift. Finally, considering tensor decomposition characteristics in data dimension transformation and data compression, a score prediction model based on the “user-item-interest group” tensor decomposition is constructed. Simultaneously, a time decay function is introduced in the construction of the tensor to dynamically fit the user behavior and further improve prediction accuracy. Experiments show that the proposed framework can effectively improve the recommendation accuracies resulting from both sparse scoring data and dynamic user interest division.