40 results about "Complex network analysis" patented technology

The invention discloses a three-network fault propagation risk assessment method and system for complex network analysis, and belongs to the technical field of power system toughness assessment and disaster risk management, and the method comprises the steps: collecting node and edge data of a power network, an information network and a traffic network, and forming a heterogeneous network topology structure; constructing a node importance evaluation function; carrying out weighted correction, and outputting a three-network coupling node importance degree sequence; generating a fault event triggering list; updating the node state matrix until the node state does not change any more, and outputting a fault influence range and a fault propagation path; and calculating a three-network overall connectivity loss rate, function recovery time estimation, key node fault sensitivity and a coupling dependence vulnerability index, and outputting a three-network fault propagation risk assessment index. According to the method, the dependency relationship and the influence strength among the three networks can be truly reflected, the contribution degree of the nodes to the system toughness is quantified, and the multi-dimensional, quantifiable and explainable effective evaluation of the fault propagation risk under the complex network is realized.

The invention discloses a public opinion propagation prediction method based on a dynamic time weighted Renyi entropy and a graph neural network, and the method comprises the steps: constructing a public opinion propagation network based on the hot topic information of a network platform; constructing a public opinion propagation prediction model based on the dynamic time weighted Renyi entropy and the graph neural network; and inputting the public opinion propagation network into the public opinion propagation prediction model to complete public opinion propagation prediction of the hot topic information. According to the method, the machine learning method, the complex network and the graph entropy theory are combined, a new view angle is provided for public opinion propagation prediction, a new thought is provided for application of the time weighted entropy features in complex network analysis, and the method has important theoretical value and practical significance.

The invention provides an entity alignment and graph fusion method and device based on a large language model, and the method comprises the steps: processing triple data of a general knowledge graph through a large language model, enabling the triple data to be consistent with a to-be-fused domain knowledge graph in format, and extracting an entity set and a relation triple set; calculating structural similarity, entity description similarity and relation description similarity among entities by utilizing the entity alignment model and the large language model; performing weighted fusion on the similarity, calculating entity similarity, merging entities meeting alignment conditions, and directly adding unaligned entities and relationships thereof into a new map; and storing fused atlas data by using a complex network analysis library to obtain a fused new atlas. Through a multi-information fusion mode, the method comprehensively considers the structure, description, relation and other features of the entity, greatly improves the accuracy of entity alignment, and improves the quality of map fusion.

The invention relates to the technical field of complex network analysis, and provides a social network overlapping community discovery method, which comprises the following steps: S1, calculating a potential influence value of each node, and screening out a set of core nodes in a network according to the influence values; s2, calculating the viewpoint convergence speed between the nodes by using the viewpoint value change in the DeGroot viewpoint dynamical model, and calculating the propagation weight between the nodes according to the average accumulated viewpoint difference value for measuring the viewpoint convergence speed; s3, constructing a membership degree vector for each node, and updating the membership degree vector in combination with the propagation weight in each round of propagation; s4, adaptively adjusting a community membership degree threshold value of label propagation according to the community network scale and the community characteristics, and allocating the community network attribution of the nodes according to the community membership degree threshold value; and S5, according to the propagation result of the initial label, carrying out community structure correction and redistribution of undistributed nodes. The defects of randomness, overlapping modeling and calculation efficiency of a traditional method are overcome.

The invention discloses an intelligent ore mining control method and system based on machine learning, and relates to the technical field of mining intelligentization, and the method comprises the steps: building a mining area geologic model based on a convolutional neural network model and a support vector machine model, inputting an ore body geologic feature map into the mining area geologic model, and generating a geologic analysis report; fusing the ore body geologic feature map and the geologic analysis report through a Bayesian updating method to obtain a geomechanical map; through a data fusion technology and a real-time analysis algorithm, a geomechanical map and mining area real-time monitoring data are combined, ore mining risk factors are analyzed, and an ore mining strategy is dynamically adjusted to generate an optimized ore mining instruction. The mining area geological environment data is converted into high-quality graph database nodes and edges, and key geological features are identified by using a complex network analysis technology, so that the ore extraction efficiency, safety and scientific decision-making capability are remarkably improved.

The invention belongs to the field of financial science and technology, and particularly relates to a financial activity public opinion data fusion system based on a knowledge graph, and the system comprises the following steps: 1, a public opinion data intelligent fusion and processing module captures public opinion data from a plurality of data sources, carries out the standardization processing, and adjusts the weight of the data sources for fusion; 2, extracting risk features in the public opinion data through natural language processing in a risk feature extraction and intelligent analysis module; 3, a knowledge graph construction and complex network analysis module constructs a knowledge graph of local illegal financial activities; 4, a deep neural network risk prediction and situation awareness module is used for training public opinion data by using a deep neural network; 5, triggering risk early warning and monitoring public opinion changes according to the deep learning model and a network analysis result; 6, a risk feedback and system optimization module collects risk feedback and adjusts model parameters; and step 7, the report generation and analysis system generates a detailed risk report according to the risk early warning result.

The invention discloses a time-space factor driven complex network anomaly detection method and device and a medium, and relates to the technical field of complex network analysis and network security. The method comprises the following steps: constructing a network security industry chain network, coding time features and spatial neighborhood features based on a time decay function to obtain a time-space factor, calculating a dynamic relation weight according to the time-space factor, obtaining a feature vector of each node based on time sequence sliding window sampling and contrast window learning, and obtaining a feature vector of each node; determining the node-level characteristic deviation of each node based on the characteristic vector of each node, and calculating the score of each node based on the node-level characteristic deviation, the structure evolution rate and the cross-network alignment anomaly of each node; and based on a dynamically set score threshold, if the score of the node exceeds the set score threshold, determining that the node has an abnormal behavior. According to the invention, the accuracy of complex network anomaly detection can be improved.

The invention provides a PageRank-based identification method and system for important nodes in a fused directed weighted network, and relates to the technical field of complex network analysis, the method comprises the following steps: constructing a fused directed weighted network, and fusing two single-layer directed weighted complex networks a and b by multiplexing partial nodes to form a fused network c; the output intensity of the node c of the fusion network is calculated according to the calculation formula that # imgabs0 # and # imgabs1 # are the output intensity of the node vi, and # imgabs2 # and # imgabs3 # are the number of the nodes, belonging to the single-layer directed weighted complex network a and the node number of the single-layer directed weighted complex network b, of neighbor nodes of the node vi; the importance value of a node v to be evaluated is iteratively calculated according to the calculation formula that # imgabs4 # sigma (0 < sigma < 1) is a damping coefficient, n is the total number of nodes of the fusion network, INR (vi) is the importance value of a node source vi pointing to the node v, and wc (vi, v) represents the directed edge weight from the node vi to the node v in the fusion network c; and according to the calculation result of the c node importance value of the fusion network, node importance sorting is carried out, and key nodes are identified.

The invention relates to the technical field of complex network analysis and data mining, and provides a scientific research cooperation analysis method for edge crossing attribute network clustering, and the method comprises the steps: obtaining original data, and carrying out the preprocessing of the original data; an undirected empowerment cooperative network is constructed and processed to obtain a core sub-graph, and the maximum connected component LCC of the core sub-graph is extracted; constructing a node multi-dimensional feature vector by using a natural language processing technology; using an ECV algorithm to determine an optimal community number; constructing a topological feature matrix and an attribute similarity matrix by adopting an ANCA algorithm, and fusing the topological feature matrix and the attribute similarity matrix; hyper-parameters of the ANCA algorithm are adjusted and optimized in a grid search mode, and community division and analysis are carried out. The method has an extremely important application value and a wide social prospect in optimizing resource configuration, evaluating team influence, promoting academic communication and cooperation and guiding development of a scientific research direction.

The invention provides a method, a device, equipment and a medium for assessing potential risks of a cavern repository, and relates to the technical field of nuclear radiation environment influence assessment, and the assessment method comprises the steps: obtaining a basic list and project feature information of cavern disposal safety assessment; screening and evaluating the basic list according to the project feature information to generate a corresponding key list; and carrying out network analysis processing on the key list through a complex network analysis algorithm to generate a corresponding relation network, and carrying out scene analysis processing on the relation network based on preset key scene parameters to generate a potential risk scene of the cavern repository. According to the method, efficient identification, scientific classification, accurate assessment screening and deep processing analysis of feature-event-process data are realized, powerful support is provided for potential risk assessment of the cavern repository, and scientificity and reliability of decision making are improved.

The invention discloses an engineering project complex network-oriented key node mining method and system, and belongs to the technical field of engineering project complex network analysis. In order to solve the problem of insufficient utilization of global topological information and node attribute information in item node importance sorting in the existing method, the invention mainly adopts a disturbance perception graph neural network and a negative sampling technology, and combines an encoder-decoder structure to carry out network reconstruction. Through multi-time disturbance and comparative analysis of the reconstructed graph, the key nodes in the engineering project network can be effectively identified and quantified, and the accuracy and reliability of node importance evaluation are improved.

This invention discloses a safety analysis method for ship SOFC systems, belonging to the field of risk prevention and control technology. It includes: identifying the causative factors of the ship SOFC system; constructing a failure fault tree and event tree model; constructing a Bow-Tie model based on the failure fault tree and event tree model; quantitatively analyzing potential causative factors to identify key factors affecting the SOFC system; combining the analyzed ship SOFC system accident causation data and collected ship fuel cell standard and specification data to form a data system and construct a ship SOFC system accident risk prevention knowledge graph; and conducting SOFC system safety analysis based on the ship SOFC system accident risk prevention knowledge graph. This invention clarifies the criticality, hierarchical relationship, and action path of each causative factor; constructs the causative factors into a complex accident causation network; and, combining the semantic information of the knowledge graph and the capabilities of complex network analysis, deeply reveals the accident causation propagation mechanism, providing a comprehensive perspective for understanding the occurrence and development of accidents.

The invention discloses a community discovery method combining local expansion and importance sequence updating rules, and aims to solve the problems of oscillation, randomness and local optimum of a traditional label propagation algorithm in a complex network. The network is subjected to preliminary community division in combination with the influence of the nodes and the weight information of the edges, and the influence of the nodes is calculated through the characteristics of degree centrality, betweenness centrality, graph embedding and the like. And node labels are updated by using asynchronous label propagation, so that the oscillation phenomenon in the traditional LPA is avoided. The obtained community is optimized through spectral clustering, the community structure is further refined from a global perspective by using a Laplacian matrix and a feature vector, and the accuracy and stability of community division are improved. According to the method, the community division precision in a large-scale and dynamic network can be effectively improved, and noise interference and random influence of a traditional algorithm are overcome. A more accurate, stable and explainable solution is provided for a community discovery task, and deep development of complex network analysis is promoted.

The application belongs to the field of financial technology, and specifically relates to a financial activity public opinion data fusion system based on a knowledge graph, steps of which are as follows: step one: an intelligent public opinion data fusion and processing module, which extracts public opinion data from multiple data sources, performs standardized processing, adjusts data source weights for fusion; step two: a natural language processing in a risk feature extraction and intelligent analysis module extracts risk features in public opinion data; step three: a knowledge graph construction and complex network analysis module constructs a knowledge graph of local illegal financial activities; step four: a deep neural network risk prediction and situation awareness module uses a deep neural network to train public opinion data; step five: according to a deep learning model and network analysis results, risk early warning is triggered and public opinion changes are monitored; step six: a risk feedback and system optimization module collects risk feedback and adjusts model parameters; and step seven: a report generation and analysis system generates a detailed risk report according to risk early warning results.

The present invention relates to a method and system for classifying characteristic information of special personnel based on graph neural networks, comprising: obtaining a special personnel behavior data set and special personnel characteristic information; performing data preprocessing based on the obtained special personnel behavior data set and special personnel characteristic information, including supplementing missing values ​​in the data set by using a median supplementation method; performing normalization operations on the numerical data of the data set; and performing numerical processing on the categorical data of the data set. The present invention provides a method for classifying characteristic information of special personnel based on graph neural networks, proposes a method for constructing a graph network based on the similarity between special personnel individuals, extracting graph network features using complex network analysis, and on this basis, aggregating neighbor features through graph neural networks to evaluate the behavior of special personnel.

The invention discloses a method and a system for analyzing a progressive instability mode of a rock block system based on a complex network. The method comprises the following steps: determining an adjacent relation between rock blocks through rough detection and fine detection of a bounding box; analyzing potential instability modes of the single block, including falling instability, single-sided slippage instability and double-sided slippage instability, and calculating corresponding safety coefficients; a potential instability network of the block system is constructed, rock blocks and potential instability modes thereof are expressed as network nodes, and the influence relation between the rock blocks is expressed as directed edges; and an unstable block and adjacent rocks thereof are identified through iteration, an actual instability network is constructed, and the instability sequence and hierarchy are determined. According to the method, the progressive instability process of the rock system can be efficiently and accurately predicted through the complex network model, and the problems that a traditional method is high in calculation cost, and instability sequence and hierarchy are difficult to reflect are solved.

The invention provides a self-adaptive routing method and system based on Motif statistics and reinforcement learning, an on-chip routing network (Network-on-Chip, NoC) provides a communication basis for a decentralized many-core chip architecture, and the performance of the NoC is closely related to the routing method; according to the adaptive routing method and system, the data communication path is dynamically adjusted according to the actual operation scene of the chip, and it can be guaranteed that an on-chip routing network can express efficient communication performance in different complex communication scenes. Real-time global communication information is processed based on a complex network analysis method Motif, a Motif analysis result and other routing related information are combined, and a path selection strategy is trained through a deep reinforcement learning algorithm. The adaptive routing algorithm based on global communication analysis has stronger adaptive ability, and can provide better data communication time delay and higher data throughput in a complex and dynamically changing scene.

The invention discloses a mathematical description method of a cascade reservoir group joint scheduling system based on a complex network. The mathematical description method comprises the following steps: step 1, determining a network type; 2, defining nodes of the cascade reservoir group complex network; 3, defining edges of the cascade reservoir group complex network; 4, defining an adjacent matrix A of the cascade reservoir group complex network; step 5, defining node attributes of the cascade reservoir group complex network and a constraint set of the node attributes; 6, defining natural incoming water as an input variable of the reservoir group system; 7, describing a hydraulic coupling relationship between the cascade reservoirs; 8, describing a state transition equation between the cascade reservoirs; according to the method, the mutual relation between the reservoirs can be more comprehensively captured, the network property between the reservoirs in a cascade reservoir group combined dispatching system is visually reflected, more analysis angles are provided for reservoir optimization dispatching, and a systematic mathematical description method is provided for complex network analysis of the reservoir groups.

The invention is suitable for the technical field of complex network analysis, and provides a community discovery method based on high-order enhancement, which comprises the following steps: formalizing a literature reference network into an undirected graph, and constructing an adjacent matrix and a node attribute matrix based on the undirected graph; constructing a DNMF core loss based on the adjacency matrix and the node attribute matrix; constructing a first-order Laplacian regular term objective function and a high-order Laplacian regular term objective function based on the adjacent matrix; constructing comparative learning loss, and fusing the DNMF core loss, the first-order Laplacian regular term objective function, the high-order Laplacian regular term objective function and the comparative learning loss to obtain a total objective function; performing optimization solution on the total objective function to obtain a topology embedding matrix and an attribute embedding matrix; and obtaining a community discovery result based on the topology embedding matrix and the attribute embedding matrix obtained by solving. According to the invention, the accuracy of document community division can be improved.

The invention relates to the technical field of complex network analysis and data mining, and provides a symbol network link prediction method based on node behavior similarity, and the method comprises the steps: 1, carrying out the preprocessing of behavior data, carrying out the time slice division of original node behavior data, and forming a discrete time sequence behavior matrix; 2, behavior joint modeling: carrying out statistics on behavior combinations of node pairs in different time slices, and generating a joint behavior probability matrix and an edge distribution probability; thirdly, a weight distribution mechanism is adopted, and a Softmax weighting strategy is introduced; fourthly, symbol mutual information is calculated, and a final behavior similarity value is calculated by combining node pair symbol labels; and 5, outputting behavior characteristics, and inputting results as characteristics into the link prediction model. According to the method, high efficiency and accuracy of link prediction are realized, and more reliable technical support is provided for applications such as a social recommendation system, online trust evaluation and user relationship mining.

The invention relates to the technical field of complex network analysis, in particular to a content structure rule and time rule fused target community discovery method and device, and the method comprises the steps: obtaining the message sending information and message sending times of a user; user content structure rules are identified from the text sending information, and the similarity of the content structure rules among the users is calculated through a Jaccard similarity coefficient; constructing a user document sending time rule matrix based on the document sending times, and calculating time rule similarity among users through a Pearson's correlation coefficient; establishing a network undirected weighted graph based on the inter-user content structure rule similarity and the inter-user time rule similarity; and performing community discovery on the network undirected weighted graph by using a spectral clustering method to obtain a community division result. According to the technical scheme, the hidden communities and the influence network in the social platform can be found, and the target community based on the user preference can be accurately positioned.

The present invention discloses a method for identifying key nodes in complex networks based on improved dynamic sensitive centrality, which relates to the technical field of complex network analysis and is used to identify key nodes with significant influence in aviation networks. Traditional methods mostly focus on analyzing the local or global impact of nodes, and do not fully consider the interactive impact between nodes. This method constructs a more accurate key node propagation impact assessment model by considering the node's own influence and the influence of neighboring nodes. Experimental results based on six actual aviation network data sets and the SIR propagation model show that IDS is significantly superior to existing traditional methods in node identification accuracy, especially on EU air‑2, where the Kendall correlation coefficient of IDS compared to DS centrality is improved by up to 105%. This method provides theoretical support for the stable operation of aviation networks and is suitable for identifying key nodes in different types of complex networks.

The present disclosure relates to the technical field of complex network analysis, and specifically to a method and device for discovering a target community that integrates content structure rules and time patterns. The method for discovering a target community that integrates content structure rules and time patterns includes: obtaining a user's posting information and number of posts; identifying the user's content structure rules from the posting information, and calculating the similarity of content structure rules between users using the Jaccard similarity coefficient; constructing a user posting time pattern matrix based on the number of posts, and calculating the similarity of time patterns between users using the Pearson correlation coefficient; establishing a network undirected weighted graph based on the similarity of content structure rules between users and the similarity of time patterns between users; and using a spectral clustering method to perform community discovery on the network undirected weighted graph to obtain community division results. The above technical solution helps to discover hidden communities and influence networks in social platforms and accurately locate target communities based on user preferences.

The invention discloses a vehicle pile network data aggregation method, and belongs to the technical field of multi-source data processing, and the method comprises the steps: carrying out the initial grid division of an urban road network, and optimizing the boundary of a grid through the combination of complex network analysis and a random walk algorithm; acquiring data in each grid and preprocessing the data; performing feature extraction on the preprocessed data by using a feature extraction network, and fusing the extracted features into grid features by using a grid feature fusion model; using a feature fusion algorithm to aggregate each grid feature and the corresponding predicted target value, and using the aggregated grid features and the corresponding predicted target values to perform joint training on the feature extraction network and the grid feature fusion model; and performing feature extraction and feature fusion by using the trained feature extraction network and the grid feature fusion model to complete the iterative aggregation of the vehicle pile data. According to the method, vehicle network collaborative interaction under multi-source data aggregation can be realized, and construction of an intelligent energy traffic network is supported.

The invention discloses an innovation path recommendation method based on complex network analysis and scientific and technological innovation force. The method comprises the following steps: acquiring multi-source technical achievement data; performing vectorization processing on text information of the result record, and constructing a technical map based on node vectors and node attributes; community detection and bridge node identification are carried out on the technical atlas to determine potential association relationships among different technical fields; generating a candidate path set containing cross-domain connection based on the community detection result and the bridging node information; scoring and sorting paths in the candidate path set, and comprehensively considering path weights, cross-domain connection degrees and length constraints; and screening out an optimal innovation path in the sorting result and outputting a recommendation result. The method can effectively reveal the potential relation between the technical fields, assists in discovering high-value innovation paths, provides decision support for scientific research planning and technical innovation, and solves the problems that in the prior art, the innovation path mining efficiency is low, and cross-field correlation is difficult to quantify.

The invention provides a financing risk monitoring method and system. The method belongs to the technical field of risk monitoring. The method comprises the following steps: acquiring multi-modal data related to financing activities, and preprocessing the acquired multi-modal data; based on the preprocessed multi-modal data, constructing a financing risk dynamic conduction network by using a graph theory and a complex network analysis technology; performing deep feature extraction on the multi-modal data in the financing risk dynamic conduction network to obtain a financing risk feature data set; performing risk type classification on the financing risk feature data set by applying a machine learning algorithm, and identifying different types of risks; a dynamic conduction network is constructed by utilizing a graph theory and a complex network technology, so that not only are association and propagation paths among various risk factors clearly presented, but also a visual analysis basis is provided for supervision and decision making; and the visualization and interpretability of the risk conduction path are enhanced.

The invention discloses a key node identification and network robustness collaborative optimization method, and belongs to the field of complex network analysis and network robustness guarantee. The method comprises the following steps: firstly, extracting a multi-scale structure and topological characteristics of nodes, constructing a supervision label fusing structural vulnerability and propagation potential, training a gated graph neural network to obtain node importance scores, and dividing the network into a core layer, a standby layer and a peripheral layer according to the scores to form layered topology; on this basis, a plurality of differentiated edge adding strategies are designed, a Q learning framework is introduced, the edge adding strategies are adaptively selected according to the network state, the network structure is iteratively optimized under the constraint of edge budget, and integrated collaboration of key node identification and topology reinforcement is realized. Compared with the prior art, the method has the advantages that the connectivity and service availability of the network under deliberate attacks are enhanced while the key node recognition precision is improved, and the method is suitable for distributed scenes such as a micro-service system, edge computing and the Internet of Things.

This invention relates to a method for modeling and blocking urban flooding disaster chains based on large language models and complex network analysis, belonging to the field of disaster prevention and mitigation. Using multi-source media text as input data, a large language model is used to automatically extract disaster events and their causal relationships. Through association rule analysis, support, confidence, and lift between events are calculated to construct a weighted directed disaster chain network, which characterizes the cascading propagation relationships of urban flooding events. Based on complex network theory, indicators such as the centrality, connectivity, and network efficiency of nodes and edges are calculated to identify key nodes and vulnerable propagation paths. A progressive node and edge removal experiment is designed to evaluate various intervention strategies and determine the optimal disaster chain blocking scheme. This invention enables automated modeling and quantitative analysis of disaster chain networks from unstructured text data, rapidly identifying key propagation links and system vulnerabilities in urban flooding, significantly improving the accuracy of urban flooding disaster chain identification, analysis efficiency, and the scientific nature of prevention and control decisions.

The invention discloses a dynamic community detection method and device, a medium and equipment, and relates to the technical field of artificial intelligence and complex network analysis. According to the method, neighbor view angle information and structure view angle information of each user node are fused, and network denoising and feature enhancement are realized through dual-channel graph convolution according to the similarity between neighbor nodes of each user node in a graph structure of each time step and the similarity between topological structures of each user node; according to the method and the device, the user nodes are extracted, then node representation and global graph representation of the user nodes are generated, and when the two-channel graph convolution process is trained, joint modeling structure consistency loss, potential implicit conflict loss and local-global mutual information loss are constructed, so that the node representation accuracy is improved, the utilization rate of multi-view information is improved, and the method and the device are suitable for popularization and application. And clustering is carried out based on the trained node representation, so that the division accuracy of communities in the dynamic social network is improved, and the calculation efficiency is considered.