52 results about "Social network analysis" patented technology

The invention discloses a public opinion event multi-mode semantic fusion modeling and abstract generation method and system, and relates to the field of natural language processing and social network analysis. Through the multi-mode semantic fusion technology, the short text understanding ability is improved, and the problems of semantic fuzziness and network language diversification are solved. Meanwhile, through a cross-window event cluster matching technology, an event evolution path with time continuity is constructed, and comprehensive capture of event dynamic characteristics is realized. Besides, the structured event abstract is automatically generated by utilizing the generative model, so that the consistency and the information density of the abstract are improved, and the actual application requirements are met. Through the innovations, the defects in the aspects of semantic comprehension, dynamic modeling and abstract generation in the prior art can be effectively overcome, a more efficient and accurate solution is provided for monitoring and analysis of public opinion events, and the method has wide application prospects in the fields of public opinion monitoring, emergency early warning, social media data analysis and the like.

According to the privacy enhancement type (k, d)-truss community search method for the social network, a novel KTG tree structure is constructed by fusing k-truss and G-tree indexes. According to the index structure, hierarchical community decomposition of a social graph and social distance information are fused, a refined boundary vector coding scheme is designed to support efficient distance calculation, and a double-cloud-server non-collusion architecture integrating improved homomorphic encryption and matrix encryption technologies is constructed. Through a two-stage security query process of first structure filtering and then distance verification, on the premise of protecting full-process privacy of a graph structure, a query intention, a distance matrix and an intermediate calculation result, efficient and accurate search of a close community in a large-scale social network is realized. The method is suitable for various scenes such as social recommendation, risk control, public opinion analysis and anti-fraud, and the problems of privacy disclosure and calculation efficiency in social network analysis are effectively solved.

The invention relates to a node injection attack method based on adaptive target selection, and the method comprises the following steps: S1, target node selection: calculating a comprehensive score of a node based on uncertainty and topology centrality, and dynamically selecting a target node set of a current attack round; s2, feature generation: using an adaptive feature generator to generate node features which are similar to target node distribution and have strong aggressiveness; and S3, disturbance edge construction: selecting an optimal disturbance edge connection mode for the injection node according to strategy network output in reinforcement learning. According to the method, the attack flexibility can be improved through dynamic target selection, the attack performance can be remarkably enhanced through combination of disturbance characteristics and structures, and the method has good concealment, expandability and generalization ability. The method is widely applied to security evaluation and defense research fields related to graph neural networks, such as social network analysis, recommendation systems, knowledge graphs and the like.

The application discloses a cross-department collaboration efficiency optimization method fused with social network analysis, comprising the following steps: multi-modal data acquisition and space-time labeling: collecting communication data, behavior data and physiological signal data in cross-department collaboration, adding time stamp and node identification to each data, and associating the node identification with department attributes and role characteristics in the social network; dynamic identification of cognitive bias: based on the social network node interaction data, adopting natural language processing and sentiment analysis technology to detect logical fallacy of the communication text, and generating bias type label and propagation intensity index in combination with physiological signal wave, the application improves collaboration efficiency, accurately identifies the cognitive bias propagation path in cross-department collaboration through multi-modal data fusion and social network analysis, dynamically intervenes to reduce invalid communication, shortens the task response and dispute resolution time, enhances collaboration resilience, and the space-time enhancement and critical state early warning mechanism of the propagation graph can avoid the risk of network structure imbalance in advance.

The invention discloses an edge graph neural network edge point modeling method based on dynamic graph optimization, and belongs to the technical field of graph neural networks and deep learning. According to the method, a graph structure containing task related nodes and auxiliary concept nodes is constructed, an iterative refining mechanism of edge point joint updating is introduced, and dynamic mutual optimization of node features and edge features is achieved. In the node updating stage, the edge features serve as weights to guide neighborhood information aggregation; in the edge updating stage, updated node features are used for recalculating edge features to form closed-loop optimization of edge guiding points and point updating edges. According to the method, the static problem of edges in a traditional graph neural network is solved, the utilization ability of the model to dynamic relations and prior knowledge is improved, and the method is suitable for complex reasoning tasks such as small sample learning, social network analysis and recommendation systems.

The invention relates to a smart home dynamic influence maximization method based on time sequence perception, and belongs to the technical field of social network analysis. The method comprises the following steps: inputting a dynamic social network data set into a time sequence snapshot generator to generate a dynamic network sequence; pre-training labeling is carried out based on an influence capacity scoring method of time perception, and dynamic influence capacity scores, static features and dynamic features of nodes in the dynamic network sequence are calculated; the dynamic network sequence passes through a space-time network model to obtain a remodeled space-time feature vector; the remodeled spatio-temporal feature vector passes through a time sequence perception enhancement module to obtain fusion features; inputting the fusion features into a candidate seed node predictor, and predicting candidate seed nodes in each snapshot; and based on the candidate seed nodes in each snapshot, selecting an optimal seed node from the snapshots through an enhanced greedy algorithm to obtain an optimal seed node set. According to the invention, the accuracy of candidate seed prediction can be improved.

The application discloses a kind of atmospheric pollution source positioning method and device based on social network analysis.The method comprises: obtaining the pollutant data of each monitoring station in a certain period in region and pre-processing;The pollutant data after pre-processing is filtered, and air pollution scene is established;Social network analysis is carried out on air pollution scene data to obtain the correlation strength between each station;According to the correlation strength between each station and the distance between each station, the intermediate centrality value of each station is calculated;According to the intermediate centrality value of each station and geographic coordinates, spatial interpolation is carried out using interpolation method, and the pollution source is positioned according to the plane data obtained after spatial interpolation.The positioning method of atmospheric pollution source proposed in the application can not only find the corresponding pollution area and key station corresponding to pollution event in complex environment, but also facilitate targeted management, and can further accurately locate the positioning range of pollution source to the vicinity of specific monitoring point, so as to efficiently and accurately locate the atmospheric pollution source.

This invention relates to the field of social network analysis technology and proposes a social network sentiment evolution model that integrates user influence and activity. To address the shortcomings of existing technologies, such as inaccurate simulations and inflexible predictions in current network sentiment analysis and sentiment evolution processes, this invention provides a social network sentiment evolution model that integrates user influence and activity. The model analysis method involves: collecting text sentiment data from target users; collecting user influence data from each user; collecting the activity level of the target social network at time t; collecting the forgetting probability of each user at time t as a measure of sentiment evolution capability; establishing a dynamic model for the sentiment evolution of the social network; optimizing the evolution probability of the information propagation model based on text sentiment; and analyzing the evolution process of the dynamic model for the sentiment evolution of the social network based on the optimized information propagation model. This model is suitable for application in social network analysis.

The invention discloses a subgraph editing distance calculation method based on a graph neural network, and relates to the field of graph calculation and artificial intelligence, and the method comprises the following steps: obtaining a query graph and a target graph; generating node-level and edge-level representations for the query graph and the target graph through a unified graph isomorphism encoder to capture the influence of node and edge specific editing operation on graph topology; inputting the target graph into a self-adaptive graph mask module, generating mask scores of nodes and edges through a gating attention mechanism in combination with query graph representation, shielding irrelevant parts and dynamically balancing graph scale differences; and performing multi-head mask processing on the target map based on the mask score to generate a plurality of candidate substructures. According to the method, asymmetry can be effectively processed, fine-grained structure differences are captured, calculation precision and efficiency are improved, and the method is suitable for scenes such as drug discovery, social network analysis and recommendation systems.

The invention relates to the technical field of social network analysis, and discloses a method for identifying important nodes in social network data, a terminal and a storage medium. The method comprises the following steps: acquiring social network data and calculating semantic similarity between posters; the method comprises the following steps: mapping users, stickers and topics into nodes, mapping attention, interactive behaviors and similarity relationships into edges, and constructing a multi-relationship heterogeneous graph; secondly, defining a meta path to capture high-order topological dependence, performing multi-hop neighborhood aggregation on nodes, and fusing different path characteristics by using a self-attention mechanism to obtain high-order structure characteristics; meanwhile, a unified semantic space is constructed based on a pre-training language model, and semantic embedding representation of nodes is extracted. And finally, the high-order structure features and the semantic features are deeply fused to generate joint representation, and scores are calculated based on the joint representation so as to accurately identify important nodes. According to the method, the hierarchy and potential influence of the nodes in the network can be comprehensively described, so that the accuracy and robustness of important node identification are improved.

The invention discloses a small sample learning method suitable for single power grid project investment execution risk management evaluation, which relates to the technical field of power grid investment management, and comprises the steps of establishing an evaluation index system, screening and extracting key indexes, and establishing a small sample learning method based on a semi-supervised prototype network. According to the method, a semi-supervised form of a prototype network is further defined on the basis of the prototype network, then typical annotations are generated by using kernel density estimation, and finally empirical analysis is carried out. Through screening and defining of multi-stage key evaluation indexes, in combination with social network analysis and an interpretation structure model, process type, driving type and result type key indexes are comprehensively identified; the semi-supervised prototype network and kernel density estimation are innovatively applied, so that risk classification and management can be effectively carried out under the condition of small sample data, and the actual problem of lack of big data is solved.

The application discloses a social network influence maximization privacy protection method based on zero-knowledge proof and belongs to the technical field of cryptography and social network analysis, and comprises the following steps: constructing a network topology structure of a social network; generating a network commitment based on a Commitment circuit according to the network topology structure; after topological sorting according to the network topology structure, obtaining a final seed node set and a final proof set by using a greedy algorithm; compressing the final proof set by using a tree-shaped proof recursive method to form a final proof; and taking a tuple formed by the final proof, the network commitment, the final seed node set and a final influence diffusion value as a social network influence maximization privacy protection result. The method can provide a verifiable influence maximization calculation result to a third party on the premise that the network structure of the social network is completely private, and ensures efficient operation on an actual social network.

The application provides a symbol graph-oriented influence maximization method and system, and the method comprises the following steps: acquiring symbol graph data, and performing representation learning on nodes by using a symbol-aware attention layer to generate symbol-aware node embedding; a seed set representation module based on a variational autoencoder is constructed, a conflict-aware regularization term is introduced to punish negative relationships in the seed set and encourage positive cooperation; a double-path diffusion model is used to simulate the propagation dynamics of positive and negative influences to obtain node net influence; the modeling capability of a complex teacher model is transferred to a lightweight student model through knowledge distillation; finally, a hybrid reasoning strategy is used to output a seed node set with high net influence and minimum internal conflict. The application can effectively process positive and negative edge heterogeneity information in the symbol graph, accurately simulate the inhibition effect in the influence propagation, generate a high-quality seed set, and has important application value in the fields of social network analysis, viral marketing and the like.

The invention discloses a microscopic cascade prediction method and system based on user portrait decoupling and multi-source signal fusion, and belongs to the field of artificial intelligence and social network analysis. The method comprises the following steps: firstly, learning user static representations in a social network and a historical cascade graph by using a graph neural network, and obtaining community-level representations through clustering; secondly, the cascade data is constructed into a multi-time granularity hypergraph sequence, and a hypergraph attention network is used for learning time sequence dynamic representation of a user; then, through a double-branch parallel architecture, individual features and common features of the user are decoupled from the static and dynamic representations respectively, and constraints are introduced to optimize feature learning; then, adaptive gating fusion is carried out on the two types of features, and unified user representation is formed; and finally, predicting a next user most likely to participate in the information cascade based on the unified representation. According to the method, the accuracy and the interpretability of micro cascade prediction are improved by decoupling the heterogeneous behavior mode of the user and fusing the multi-source information.

A method of streamlined selection and application of a makeup look involving personalized makeup look recommendations with respective step-by-step makeup routine instructions for executing a selected makeup look. A predictive model identifies one or more makeup look recommendations from a database of makeup looks to present to a user. Recommendations are generated from a combination of user demographic data, user-input preferences, social networking analytics (e.g., frequently used makeup looks among users such as “trending” makeup looks), and user behavioral history data obtained from a smart device application and a smart cosmetic device system. Each respective makeup look can be executed from a respective detailed tutorial guide comprising step-by-step instructions on how to apply cosmetic substances to achieve the respective makeup look.

The invention relates to a graph enhancement method and system based on core subgraph retention and structure compression and a storage medium, and relates to the technical field of graph contrast learning. According to the method, a graph augmentation strategy combining core subgraph extraction and information compression is provided for solving the problems that in existing graph contrast learning, random augmentation is prone to damaging a semantic structure, and redundant information is introduced. The method comprises the following steps: encoding an input graph by using a flexible connection GNN, and identifying and retaining a core sub-graph with a stable structure and dense semantics through adaptive frequency spectrum filtering; and an enhanced view is generated on the basis of the core subgraph, semantic consistency is kept by using comparison consistency loss, and compression loss based on an information bottleneck principle is introduced to remove redundant information. According to the method, noise can be effectively suppressed while the semantic core of the graph is reserved, the discrimination and generalization performance of representation are improved, and the method is suitable for various graph representation learning tasks such as molecular modeling, social network analysis and recommendation systems.

The invention relates to a dynamic authority allocation method and device for cross-department knowledge sharing and security isolation, computer equipment, a storage medium and a computer program product. The method comprises the steps that user access logs are collected and preprocessed, modeling is conducted on a user behavior sequence through a recurrent neural network model for the preprocessed user access logs, and a behavior analysis result is obtained; constructing an enterprise knowledge graph in combination with natural language processing and social network analysis technologies, and generating a knowledge demand evaluation result according to the enterprise knowledge graph; determining an optimal permission scheme in combination with a behavior analysis result and a knowledge demand evaluation result; according to the user access context and the behavior analysis result, matching an adaptive security isolation strategy from a preset strategy library; and executing real-time permission updating according to the optimal permission scheme and the security isolation strategy. By adopting the method, the management cost and the compliance risk can be reduced, and the method has high expandability and scene adaptability.

The invention relates to a marketing topic key user and influence measurement method based on multi-feature fusion, and belongs to the technical field of social network analysis. According to the method, modeling is performed on text data through an LDA topic model, and meanwhile, a marketing topic set is extracted by utilizing a clustering algorithm. The topic correlation is used as an index for measuring the correlation degree between the text and the topic, and the quality and correlation of the text and the topic are effectively evaluated. Secondly, introducing a fuzzy mathematical theory, and establishing a fuzzy comprehensive evaluation model based on information entropy; various characteristics of the user are fused as evaluation indexes, and the weight is determined by using an entropy weight method, so that the interference of subjective preference is reduced. And finally, constructing a multi-dimensional heterogeneous network based on marketing topics. By fusing the attribute characteristics of each node and utilizing the random walk strategy, the user behavior and interaction information are comprehensively considered, so that the influence of the user in the marketing topic is more comprehensively evaluated.

The invention relates to a hot event propagation path prediction method and system based on multi-modal big data. The method comprises the following steps: based on a preset trigger, identifying seed content from a content publishing data stream, and acquiring propagation behavior data by taking the seed content as a center; based on the seed content and the propagation behavior data, constructing a preliminary propagation network structure to obtain an early propagation network; analyzing a propagation chain in the early propagation network to obtain a propagation chain analysis result, and based on the propagation chain analysis result, evaluating a user tendency response behavior to obtain a response tendency prediction matrix; and performing propagation simulation based on the early propagation network, the propagation chain analysis result and the response tendency prediction matrix to obtain propagation probability distribution. By adopting the method, the accuracy of propagation prediction can be improved by utilizing multi-modal data and social network analysis, the combination of content semantics and user behaviors is emphasized, and the method is suitable for the hotspot event propagation analysis of platforms such as social media and the like.

This invention relates to a time-aware method for maximizing the dynamic influence of smart homes, belonging to the field of social network analysis technology. It includes the following steps: inputting a dynamic social network dataset into a time-series snapshot generator to generate a dynamic network sequence; pre-training and labeling using a time-aware influence capacity scoring method to calculate the dynamic influence capacity score, static features, and dynamic features of nodes in the dynamic network sequence; passing the dynamic network sequence through a spatiotemporal network model to obtain a reshaped spatiotemporal feature vector; passing the reshaped spatiotemporal feature vector through a time-aware enhancement module to obtain fused features; inputting the fused features into a candidate seed node predictor to predict candidate seed nodes in each snapshot; and selecting the optimal seed node from the snapshots using an enhanced greedy algorithm based on the candidate seed nodes in each snapshot, obtaining the optimal seed node set. This invention can improve the accuracy of candidate seed prediction.

The application provides a subgraph-based information cascade prediction method and system, belonging to the technical field of social network analysis and neural network, comprising the following steps: S1, constructing a deep learning information cascade prediction model CasSubTS, inputting collected user published information through an input layer, and constructing an information cascade graph G; S2, inputting G into a subgraph sampling layer, dividing G into a plurality of information cascade subgraphs according to different time steps, and converting the information cascade subgraphs into adjacency matrices; aggregating node features of the adjacency matrices to obtain a feature representation matrix B; S3, inputting B into a feature learning layer to obtain a characteristic vector with structural features and time sequence features; S4, inputting an input feature weighting layer, and utilizing a channel attention mechanism to perform weighted fusion to obtain a weighted characteristic vector; and S5, inputting an input prediction layer to predict a final macro cascade increment. The method can effectively predict information cascade in a social network.

This invention proposes an influence blocking maximization algorithm based on overlapping interest community detection. It includes a multi-factor competitive propagation model and an algorithm based on this model and overlapping interest community detection, used for modeling the competitive propagation process of multiple information in social networks and selecting positive information seed nodes. Modeling the competitive propagation process and maximizing influence blocking are hot topics in social network analysis. Traditional models, based on IC and LT, are difficult to accurately depict the real-world propagation process; traditional community-based algorithms divide communities based on topological structure, ignoring user interests and preferences. This invention comprehensively considers user interests, information interaction latency between users, and trust levels, enabling better modeling of the competitive propagation process; it extends the traditional community concept to interest communities, improving community detection accuracy; and it guides the selection of positive information seed nodes through the overlapping interest community structure, achieving the desired negative information influence blocking effect in a short time.

The present application relates to a kind of financial case fund flow direction traceability analysis method, belong to computer network technical field.The present application combines social network analysis technology with graph neural network technology, the advantage of both is utilized to analyze and track the fund flow direction in financial transaction network.Social network analysis identifies key nodes and group structure in transaction network by calculating node centrality and community discovery;Graph neural network analysis further extracts complex transaction features and patterns by deep learning model, and identifies abnormal transaction behavior.In addition, by utilizing advanced graph neural network models such as graph convolution network and graph attention network, combined with parallel computing and distributed training technology, the model training process is efficient and has good scalability, which meets the processing needs of large-scale financial transaction data.

The invention provides a social network user personality detection method based on multi-view feature fusion, and the method comprises the following specific steps: firstly, extracting semantic features of a posted text released by a user based on a pre-training language model, and carrying out the sentiment analysis of the posted text in combination with an external sentiment knowledge base, so as to form initial posted text embedding; thirdly, constructing a poster-personality trait heterogeneous graph, setting node interaction of three levels, and extracting multi-level interaction features by using a graph attention mechanism; furthermore, a multi-instance gating loop unit is introduced to model a user post sequence, time sequence features under different personality traits dimensions are captured, and information sharing among personality traits is promoted through an interactive attention layer. And finally, fusing the disordered interaction features and the sequence features, predicting personality characteristics of the user in multiple dimensions, and forming a comprehensive personality portrait. According to the method, the personality characteristics of the user can be comprehensively described, and the method is suitable for scenes such as personalized recommendation, user portrait construction, social network analysis and psychological health auxiliary diagnosis.

The invention provides a social relationship intimacy calculation method and system under social media privacy propagation, and relates to the technical field of social network analysis, and the method comprises the steps: building a directed weighted graph based on user privacy propagation behaviors, taking a user as a node, taking a directed propagation relationship as an edge, and fusing an edge weight with a propagation frequency, a content average sensitivity and a topological distance adjustment factor; secondly, a Personalized PageRank algorithm is improved, a path distance Gaussian attenuation transfer matrix is introduced, and a global trust matrix is iteratively calculated; a double-exponential attenuation function is adopted for historical interaction, short-term fast attenuation and long-term slow attenuation are modeled, and a time-varying weight is obtained; quantifying privacy content sensitivity in combination with information entropy and a propagation risk model; identifying a user community and quantifying a community contribution degree; global trust, time-varying weight and sensitivity multiplication are fused, normalization is carried out after community contribution degree correction, and a final intimacy score is generated. According to the method, the relationship intimacy under privacy propagation is accurately described, and dynamic evaluation is supported.

The invention discloses a competitive overlapping community influence maximization method, which comprises the following steps of: constructing a competitive independent cascade expansion model, introducing an inter-node activation probability calculation mechanism, and comprehensively considering node similarity, interaction strength and user interest preference to realize multi-information competitive propagation modeling; meanwhile, based on an overlapping community structure and reverse reachable set sampling, a node comprehensive influence evaluation method is provided, and high-influence nodes are accurately identified in combination with the number of third-order neighbors and the community overlapping degree; through important community screening and a seed budget allocation strategy, seed node selection is optimized, and the information spreading efficiency is improved; the method is remarkably superior to an existing algorithm on the aspect of a real social network data set, shows excellent performance in the aspects of influence propagation range and calculation efficiency, and is suitable for the fields of social network analysis, information popularization, recommendation systems and the like.

The application provides a method and system for mining common neighbor relationship of massive data based on graph calculation, wherein the method comprises the following steps: step 1, constructing a concept metadata system; step 2, collecting data from an information system to be mined, and obtaining instances of concepts according to the concept metadata system; step 3, uniformly encoding attributes in the instances; step 4, storing the encoding and the corresponding instances in a graph database in a distributed storage mode; and step 5, relationship mining, extracting instances from the distributed graph storage, discovering common neighbors between the instances according to the characteristics of the encoding, filtering according to preset rules, and generating recommended relationships; and the system is used for realizing the above method. The application can be widely applied to the fields of social network analysis and recommendation system optimization, and provides an efficient solution for relationship mining of massive data.

The present application relates to an unsupervised heterophilic edge graph analysis model using an edge discriminator and a multi-channel encoder, and a learning method using the same, and generates a feature information-based representation, a connection information-based representation and a weighted graph-based representation. Then, these representations are combined to generate a final node representation vector, and the consistency of the representation is enhanced through contrastive learning. Compared with the conventional single channel model, the node representation power is increased, which leads to excellent node classification accuracy. In addition, the model of the present application may be applied to unlabeled datasets through unsupervised learning, and particularly exhibits excellent performance on heterophilic edge graph. The present application may be applied to data analysis with complex relationships in the fields such as social network analysis, recommendation systems, and bioinformatics, and may be utilized for various graph-based tasks such as anomaly detection and link prediction.

The invention provides a multilayer signal interactive propagation method, device and equipment used in a multilayer system, and relates to the technical field of social network analysis and information propagation. Comprising the following steps: acquiring basic information and to-be-propagated information of a multi-layer complex system; constructing a multi-layer hypergraph structure according to the basic information of the multi-layer complex system; determining an associated hyperedge set of each node in the common node set; performing iterative screening operation from the common node set according to the initial coupling degree to obtain a seed node set; performing intra-layer propagation on the to-be-propagated information in the current hypergraph layer according to the initial propagation state and a preset intra-layer propagation rule; and when it is detected that propagation of the to-be-propagated information in the current hypergraph layer is completed, jumping the to-be-propagated information to a next hypergraph layer according to a preset cross-layer propagation rule for intra-layer propagation until no newly added activation node exists. According to the method provided by the invention, efficient and accurate multi-layer influence propagation modeling and seed optimization are realized, and the authenticity of signal propagation simulation and the time sequence accuracy are improved.

The application discloses a kind of social group member dynamic management method and system, it is related to social network analysis field, including, the online interaction data and offline co-occurrence data of social group member are acquired, and are washed, the online interaction data and offline co-occurrence data after washing are converted into structured data set;Structured data set is imported into graph database to obtain initial weight, and dynamic attenuation weight is calculated, and basic relationship network is constructed;The edge weight matrix in basic relationship network is converted into quantum bit entanglement state, quantum bit entanglement state is measured to form quantum probability amplitude, quantum entanglement determination threshold is set, and quantum relationship thermodynamic diagram is generated by Grover algorithm.The application realizes the accurate management and risk prevention and control of social group member relationship by the technical scheme that quantum computation is combined with dynamic entropy analysis, the edge weight matrix of basic relationship network is converted into quantum bit entanglement state, and quantum relationship thermodynamic diagram is generated using quantum measurement and Grover algorithm.