54 results about "Complex network analysis" patented technology

The invention belongs to the field of data mining, aims to solve the problem of excavating core authors in an academic field and intelligently extracting related information of the authors, and provides an improved academic core author excavation and information extraction method and system based on a core node discovery algorithm in the social network analysis technology. The method combines the vertical search technology, the social network analysis technology and the text analysis technology, and can find the core authors or groups of the academic field in the mass of information to further obtain related information of the authors. The method uses the vertical search technology to collect open source literature data, uses the bibliometric technology and the complex network analysis technology to analyze the importance of a variety of social entities in the data, and utilizes a community discovery algorithm to perform entity clustering based on the closeness degree of relationships between entities to find out an academic community. Users can find the core authors or an institution according to an entity importance ranking, and find the leadership team according to published articles amount distribution of cooperative groups.

The invention provides an electroencephalogram signal characteristic extracting method. Network average route lengths and clustering coefficients are calculated through wavelet reconstruction, windowing horizontal visibility map complex network conversion and complex network analysis. The average route lengths and clustering coefficients composed of electroencephalogram signals are calculated to achieve characteristic analysis of electroencephalogram signals and chaotic time sequence signals of the electroencephalogram signals of different rhythms. The electroencephalogram signal characteristic extracting method has the advantages that one-dimensional chaotic time sequences are converted into complex networks, according to analysis of network characteristic parameters, fractal characters of the electroencephalogram signals are revealed, the complex non-linearity signals of the electroencephalogram signals are depicted from a brand new angle. The electroencephalogram signal characteristics can be applied to automatic diagnosis of mental disease and a characteristic identifying module of a brain-machine port system. The electroencephalogram signal characteristic extracting method can effectively distinguish the electroencephalogram signals of an epilepsia attach stage and an epilepsia non-attach stage, the equation p<0.1 is met after Mann-Whitney detection, and the electroencephalogram signal characteristic extracting method can be applied to epilepsia electroencephalogram automatic identification.

The invention relates to a function call graph key node recognition and identification method based on complex network analysis. The method comprises the steps of carrying out lexical analysis, grammatical analysis and control stream analysis through a sound code static analysis technology, thereby obtaining source code call relationship data and function length data, and storing all data in a database in a classification mode; generating a call graph according to the obtained data, calculating node indexes of the call graph by employing a complex network analysis method, wherein the indexes comprise interaction degree, closeness centrality, node betweenness and function length, and calculating node key degree by employing a multi-attribute decision-making method; and through combination of user demands and state data stored in the database, calculating a key node rank and corresponding grey proportion data by employing the obtained node key degree data of the function call graph, and carrying out visual display. According to the method, the key nodes in the complex call relationship graph can be recognized and identified rapidly, and the working efficiency is remarkably improved.

The invention discloses a protein complex recognizing method based on multi-source data fusion and multi-target optimization, comprising: preprocessing protein interaction network data to obtain adjacent matrixes; primarily clustering protein complexes to obtain an initial protein complex module; further optimizing the initial protein complex module, fusing topological structural features of the protein interaction network data and functional similar features of GO (gene ontology) annotation data during optimizing, and performing optimizing operation in conjunction with an adaptive multi-target blackhole optimization algorithm to obtain a more precise protein complex module; postprocessing to obtain a final optimal protein complex. The method of the invention has the advantages that protein complex recognition speed and precision are increased, the method is applicable to protein interaction networks and extensible to the analysis of other complex community networks, and the method isvery practical in complex network analysis.

The invention discloses a network community discovery method, which is used for community structure discovery in complex network analysis. The entire community discovery process of the present invention is divided into two stages. In the first stage, the node subgraph of each node in the network is used to divide the community into the community; in the second stage, based on the node subgraph The community partition results determine the order of node label propagation and update rules of each node community label. When propagating and updating the community labels of nodes in the network, the propagation order is no longer random, but the information entropy of nodes is used to measure the amount of information of network nodes, and the labels are propagated in the order of information entropy from small to large, so that Avoid the reduction of accuracy caused by randomness and uncertainty; at the same time, the update principle of node labels is no longer simply dependent on the frequency of labels, but comprehensively evaluates the uncertainty of node labels using optimal modularity and information entropy , thus avoiding the blindness of label updating.

The invention discloses a multi-mode depth network embedding method for fusing structure and attribute information, and relates to the technical field of complex network analysis. The method comprisesthe steps of establishing a network adjacency matrix and an attribute matrix, performing preprocessing, serially inputting structural features and attribute features into an encoder and a decoder, outputting a reconstructed adjacency matrix and an attribute matrix, updating a parameter iterative calculation and the like, and finally taking the output of the encoder as a final node representation.Based on the depth learning method, the invention can overcome the shortcoming that the existing shallow linear method is difficult to depict the highly nonlinear structure of the network, can map nodes in the network to a low-dimensional embedded space, and effectively maintains the structural characteristics and attribute characteristics of the nodes.

The invention discloses a brain network topology difference fast extracting method based on network synchronicity. By the method, work load and time can be reduced evidently, difference of brain network topology structures can be found fast, the difference connecting map can be built fast and accurately, and a 64-lead brain functional network can be completed in a few seconds. Synchronicity coefficients of brainwave lead time sequences are calculated to build a synchronicity coefficient matrix, the synchronicity coefficient matrix is converted into a binary network by using the self feature threshold of the brain network, and the topology structure difference connecting map is built by chi-square test and cross verification, and accordingly complex network analyzing can be reduced, experiment time and cost can be shortened, and the different connections are evident by comparing the topology structure difference connecting map with traditional network difference connecting maps to facilitate observation.

A wind field characterization method with empirical mode decomposition noise reduction and complex network analysis combined comprises the steps of performing empirical mode decomposition on an input wind field signal; performing phase-space reconstruction on intrinsic mode function components obtained after decomposition, and calculating predictable intensity; selecting the minimum value of a predictable intensity and intrinsic mode function order change relation curve as a demarcation point, and accumulating the intrinsic mode function components with the order greater than the demarcation point so as to obtain a noise reduction processed signal; performing phase-space reconstruction again on the noise reduction processed signal, calculating a complex network connection matrix, and constructing a wind field time sequence complex network; and repeating the process for wind field signals of other space points in a partial wind field, and obtaining global efficiency-modularity and assortativity coefficient-global efficiency combined feature characterization of different space points. According to the wind field characterization method, noise and original signals can be distinguished accurately, and noise reduction performance is outstanding. Characterization discrimination of combined features of the complex network is obvious, and the wind field characterization method can be widely used in various fields such as meteorology, agriculture, energy and environmental protection.

The invention belongs to the technical field of complex network analysis, and particularly relates to a method and a device for detecting aging stability communities in a time-varying network. The method provided by the invention comprises the following steps: acquiring the change condition of connection between nodes along with time, and constructing a time-varying network; Using the volatility to quantitatively describe the dynamic change degree of the connection edge in the network; Initializing a community structure, and calculating the dynamic modularity of the time-varying network in combination with the fluctuation rate; And optimizing the dynamic modularity, wherein the community structure corresponding to the maximum value is the time-varying stable community of the time-varying network. According to the invention, the network dynamic change is quantitatively described by using the volatility; the network dynamic change characteristics are combined with a community detection method; The method for identifying the stable community in the time-varying network is provided, the accuracy and reliability of community detection are improved, the method has wide application prospects in different fields of social networks, biological networks, traffic networks and the like, and meanwhile a new visual angle is provided for understanding the functions and the dynamic process ofthe community structure in a complex system in actual life.

The invention discloses a text-enhancing network expression learning method, and relates to a complex network analysis technology. A new text-information-enhancing network expression learning method is put forward on the basis of non-negative matrix decomposition, and for the network structure, in combination with the first-order and second-order similarity between nodes, network expression is obtained through a decomposition similarity matrix; for the text clustering structure related to the nodes, decomposition is conducted on a text-lexical item matrix to obtain a text clustering affinity matrix, the consistency relationship is established between the network expression and the text clustering structure through the text clustering affinity matrix, and therefore network expression learning is controlled by the network structure and the text clustering structure related to the joints. By means of the method, the network structure is depicted, the text clustering structure related to the joints is also depicted, additional information beside the network structure is added to the network expression learning, the learned nodes are used for expressing more available information, and higher identifiability is achieved.

The invention provides a regional traffic signal control effect evaluation method based on a weighted complex network, comprising the following steps of 10, integrally modeling a road traffic condition and a controllable traffic signal by using a primitive law and by using the weighted complex network, wherein an intersection is a node, a road segment is a side, each side has two attributes including a traffic load and a road segment capacity, the traffic load is the existing traffic flow on the road segment and represents the demand of the traffic flow for the road segment, and the road segment capacity is the value of a road traffic capacity; 20, calculating the supply and demand matching degree of the local traffic flow-signal in the road segment in real time so as to obtain the local supply and demand matching degree; and 30, establishing a global supply and demand matching degree by using a complex network analysis method according to the local supply and demand matching degree. The invention provides an evaluation method for dynamically measuring the supply and demand matching degree of the traffic flow-signal, which can be used independently for evaluating a regional (such as a city) traffic signal control effect, and also for providing feedback information for regional traffic signal control.

The invention discloses a visual question answering problem solving method based on a complex network analysis method, which includes semantic concept network construction, nonrandom depth walk, imageand text feature fusion and classifier, Semantic Concept Networks (SCNs) are designed to mine co-occurrence patterns of concepts to enhance semantic representation, non-random depth walk realizes themapping of complex networks to low-dimensional features, On the basis of constructing image semantic concept network, depth walk algorithm is used to learn the potential relationship of nodes in semantic concept network, and the nodes in complex network are mapped to a low-dimensional feature vector, and polynomial logistic regression is used to fuse image and text features to solve the visual question answering problem. The invention excavates the concept symbiosis mode and the hierarchical structure of the cluster concept, effectively integrates the visual and semantic features of the image, as well as the natural language features, and provides a feasible way for solving the visual question answering problem.

The invention provides an evaluation method for a low-cost collection and transportation system of domestic garbage in villages and towns. The method comprises the steps that the basic data of each facility in the collection and transportation system to be evaluated are acquired; a complex network model is constructed based on the basic data; an indicator system is constructed based on the complexnetwork model; the indicator system is used to evaluate the collection and transportation system to be evaluated to acquire an evaluation result; by taking the collection and transportation system ofdomestic garbage in villages and towns as the research object, a complex network analysis method is used to describe the operation mechanism of the domestic garbage collection and transportation system; a cost evaluation indicator is constructed; the constitutive characteristic of the collection and transportation system is clearly identified; a facility type and a layout form, which control thecost of the collection and transportation system are identified; and layout form value judgment and targeted planning guidelines of the collection and transportation system are proposed. The scientific method is constructed to reduce the operating cost of the collection and transportation system of domestic garbage in villages and towns and alleviate the problems of expensive handling and difficult handling of domestic garbage in villages and towns.

The invention provides a method and system for identifying atmospheric pollution transmission key nodes based on a complex network. The method comprises the following steps of: gridding and abstracting a preset area into the complex network; an HYSPLIT model outputs the flow of the airflow track in a preset area to obtain an adjacent matrix of the airflow track weight; a pollutant concentration interpolation method is adopted to obtain an adjacent matrix of a pollution transmission weight according to the pollutant concentration of the atmospheric pollutant monitoring point position in the preset area; the adjacency matrix of the pollution transmission weight is added to the adjacency matrix of the airflow track weight to obtain an adjacency matrix of the airflow track weight and the pollution transmission weight; an in-degree strength value and an out-degree strength value of each node are calculated by utilizing an improved PageRank algorithm; and regional pollution and governance key points are identified according to the in-degree intensity value and the out-degree intensity value. According to the method, the nodes are established after the specific area is gridded, and the specific area is averagely and comprehensively abstracted, and then complex network analysis is carried out, and the method is comprehensive and accurate.

The present invention discloses a system and a method for complex brain disease targeted combination treatment analysis. According to the system and the method, a modern data mining technology, a complex network analysis technology, a cloud computing technology and a large data technology are effectively combined; potential drug targets of diseases are dynamically queried, analyzed and computed according to different brain diseases; and problems in current medical research are solved, such as long data processing time, a long research period, lack of systematicness, a single research target and the like. According to the system and the method, an authoritative medical database abroad is combined and disease treatment method research is enabled to be systematic, processed and normalized. Besides, according to the system and the method, cost of medical disease treatment analysis and research is also reduced, and application of an advanced computer technology in the medical field is expanded. The present invention aims to establish brand-new network pharmacology study method and technology, is of special significance to breakthrough in brain disease prevention and treatment research, and takes positive efforts and does contribution for supporting and leading research, development and industrialization of innovative drug in China.

The invention provides a risk identification method based on complex network analysis, and the method comprises the steps: A, finding out all nodes which need to judge the risk conditions, and building a time sequence feature for each node; b, calculating the correlation of every two points in all the nodes, establishing a risk backbone network according to the correlation result, and calculatingindexes such as network topology characteristics; c, performing classification prediction on risk events by adopting a machine learning related method based on the risk backbone network characteristics obtained in the previous step. Risk identification based on complex network analysis is completed through the above three steps. The method is high in universality, high in objectivity and easy to operate, and the problem that risks in life are difficult to measure and evaluate objectively is solved.

The invention discloses an infant three-dimensional spontaneous movement intelligent evaluation system based on a complex network. The intelligent evaluation system comprises a three-dimensional somatosensory information input module, a movement information feature extraction module, a computer with a complex network analysis module and a movement quality output module, wherein the output of the three-dimensional somatosensory information input module is connected with the input of the motion information feature extraction module, the output of the motion feature extraction module is connected with the input of the computer with the complex network analysis module, and the output of the computer with the complex network analysis module is connected with the input of the movement quality output module; according to the system, the movement characteristics of the infant can be comprehensively reflected, targeted complexity characteristic evaluation is carried out on spontaneous movement of the infant, and large-scale and intelligent popularization and application of infant cerebral palsy screening are met.

The invention provides a robot intelligent idea control method based on a modal migration complex network. The method comprises the steps of through picture acquisition equipment carried by each robotin robots, obtaining surrounding environment information, so that the robot has a target recognition function; acquiring a four-electrode SSVEP electroencephalogram induced when a testee stares at aflicking picture in a visual excitation interface and uploading the four-electrode SSVEP electroencephalogram to an upper computer through WiFi wirelessly; using a polybasic experience modal decomposing method to process the obtained four-electrode SSVEP electroencephalogram, through the combination of a modal migration complex network analysis theory, accurately classifying the four-electrode SSVEP electroencephalogram, reversely deducting the visual excitation picture stared by the testee, then generating a robot formation control instruction and achieving robot intelligent idea control. Direction targets capable of being selected by the robot are richer, and the analyzing and processing capability of the signal and the recognition control accuracy are high.

The present invention relates to a complex network overlapping community discovery method, and belongs to the technical field of complex network analysis, in order to solve the problem of the discovery of a community structure with overlapping characteristics in a complex network. The method comprises the main steps of: representing a complex network as a form of a graph, and using nodes and edgesin the graph to describe the network; calculating a connection factor of each network node in the graph; calculating a separation factor of each network node in the graph; calculating the representativeness of each network node in the graph; sorting the nodes in the network according to the connection factors, and selecting a leader node of the network community in the nodes; initializing the community membership degree of the leader node; according to the node connection factor and the similarity, calculating the membership degree of the non-leader nodes about each network community throughthe recursive process; and outputting an overlapping community discovery result. The method disclosed by the present invention can be used to obtain a reasonable and reliable complex network overlapping community discovery result.

The invention discloses a method and system for mining and searching unstructured text data in a financial field and provides a scheme of named entity identification for the unstructured text data, association relationship mining between entities and association relationship network construction and utilization in the financial field, and the scheme is mainly for information extraction in the financial field, named entity identification, association relationship mining and the construction and utilization of an association network. The system comprises a data acquisition module, a data cleaning module, a data preprocessing module, an association mining module, an association map construction module and a complex network analysis module. The invention can complete the basic data analysis and information extraction work, construct an economic map by using the mined information and use the economic map to mine deep information and a hidden association, so that the data have more intuitivereadability and utilization value.

The invention relates to a complex network analysis and spatial effect evaluation method. The method comprises the steps: building an urban road network, carrying out the centrality analysis, and reflecting the spatial effect through evaluating the impact on the rental price of a residential district from a centrality index; constructing a rail transit and ground bus overlay network based on the actual running time of rail transit and ground buses, performing centrality analysis, and reflecting the spatial effect of the rail transit and ground bus overlay composite network by evaluating the influence of centrality indexes of the rail transit and ground bus overlay network on the rental price of the residential district; comparing the evaluation results of the centrality indexes of the urban road network and the rail transit and ground bus overlay network, and respectively comparing the influence degrees of the centrality of the urban road network and the centrality of the rail transitand ground bus overlay network; and comprehensively analyzing the influence of the rental price of the residence community, and comprehensively evaluating the space effect generated by the rental price. According to the invention, the accuracy of complex network analysis and spatial effect evaluation can be improved.

The invention discloses a public space vitality measurement method and system based on trajectory positioning data. A method of combining spatial statistics and complex network analysis is adopted; massive crowd trajectory positioning data are cleaned and aggregated; the spatial stroll data set is automatically extracted and constructed, and then the integration of the revisit rate, the activity mixing degree and the network centrality is adopted as the measurement index of the public space vitality, so that the defects of the existing evaluation system are overcome, the deviation caused by directly measuring the vitality by using the scale density is avoided, and meanwhile, the real-time measurement can be realized.

The invention discloses a method and a system for judging relevance between integrated circuit performance and complex network characteristics thereof. The method comprises the following steps: carrying out multi-tool physical design on a super-large-scale integrated circuit in a layout and wiring stage in the physical design, i.e., carrying out layout and wiring on an initial circuit by using different tools to obtain completely different layout diagrams and layouts so as to obtain different circuit performances; afterwards, complex network conversion is conducted on the layout diagram and the layout of the circuit, feature parameters of a complex network diagram of the circuit are calculated through a complex network analysis tool, network feature parameter-circuit performance correlation coefficients are calculated according to circuit performance changes and feature parameter changes, and the correlation between the performance of the integrated circuit and the feature parameters of the complex network is judged. According to the invention, the circuit performance and characteristics are not changed in the process of carrying out complex network conversion on the layout diagramand the layout, and the method has transparency.

The invention discloses a method for dividing a network of academic cooperation authors, belonging to the technical field of complex network analysis, wherein the method comprises the following steps:1) downloading paper publication data of a certain discipline or a certain period of time from an academic social network or an academic journal; 2) constructing collaborative author network, whereinin the downloaded paper publishing information, all the authors are the vertices of the co-author network after de-duplication, if two authors have co-published papers, then one edge should connect the vertices represented by the two authors, and the number of papers published by the two authors should be the weight between the two vertices. 3) dividing a collaborative author network according tothe improved c-means algorithm, the improved algorithm having the good clustering performance for the unbalanced data set; 4) outputting an academic community dividing result. The method can mine anacademic community with a small number of vertices in a large-scale collaborative author network.

The present invention discloses a task-state electroencephalogram signal analysis method based on algebraic topology and belongs to specific applications of a complex network analysis technology in the field of neural signal processing. The method comprises the following steps: using electroencephalogram signals as a data source, constructing distance relationship of electrodes at different spatial positions by calculating coherence between the electrodes, using the algebraic topology method to dynamically construct a simplex-based brain functional network, characterizing task-state electroencephalogram signal neural characteristics, further analyzing nature of the brain functional network by calculating Betti number and Euler's characteristic number, and realizing a quantitative study ofa brain function model of subjects under a task state. The task-state electroencephalogram signal analysis method is verified to perform well in the task-state electroencephalogram signal analysis, provides the new method for measuring neural responses in the task state, explores new rules and evidence for brain-like computing, thus can inspire artificial intelligence frameworks and specific algorithms design, etc., and promotes development of a new generation of artificial intelligence.

The invention discloses a blockchain digital currency identification method and device, electronic equipment and a storage medium, and belongs to the technical field of digital currencies, and the method comprises the steps: obtaining transaction data of digital currencies, carrying out the clustering of the transaction data according to a transaction address, drawing a digital currency transaction graph, and dividing the digital currency transaction graph into a plurality of communities according to communities; calculating a network analysis graph index of each community by adopting a complex network analysis method; obtaining a transaction graph structure of the known digital currency, and calculating a network analysis graph index of the transaction graph structure of the known digital currency; combining all the communities and the transaction graph structures of the known digital currencies to form a data set; and identifying the data set, and identifying a transaction graph structure in the data set. According to the invention, the transaction graph structure and the transaction mode of the digital currency can be identified.

The invention discloses a method for extracting plane geometric indexes and topological structures of rivers. The IPC classification number of the method is E02B1 / 00. The method is based on a complex network analysis technology, takes a satellite picture as an input quantity, and is composed of four main parts of extraction of a river channel water body, description of a river center line skeleton, identification and detection of river branching points and end points, and extraction of a river network topology structure and plane geometric indexes. The continuity of the extracted river channel is ensured as much as possible while the influence of a sand dam and the like on the plane form of the river channel is avoided, so that the method also has relatively high applicability to braided rivers, deformed river bays and the like with complex plane forms. A large amount of manpower and material resources cost brought by field measurement is saved, the accuracy and diversity of the obtained data are considered, and technical support is provided for river research.

The invention discloses a fusion topology and content community detection method based on a deep neural network, and belongs to the technical field of complex network analysis. The method comprises the following steps: mining a community structure in a network data set with content information, and respectively modeling topology and content by using modularity maximization and standardized cutting; on the basis of spectral matrix eigenvalue decomposition, matrix low-rank fitting and automatic encoder reconstruction, reconstructing theoretically similar and seamless fusion topology and content so as to construct a community detection model based on an automatic encoder frame deep neural network; and finally, using an evaluation algorithm to normalize the mutual information entropy and the Jaccard coefficient to evaluate the effectiveness of the model. The method has the beneficial effects that the topology and the content are seamlessly fused by utilizing an automatic encoder framework; and on the other hand, the network representation obtained based on the deep neural network has good community detection capability.

The invention relates to the technical field of complex network analysis, in particular to a longest circle rapid detection method and system based on an ant colony algorithm and a storage medium, and the method comprises the steps: carrying out the preprocessing of an undirected graph and a graph reduction strategy to reduce the scale of the graph, and improving the search efficiency through employing a PD-LS search strategy combined with the ant colony algorithm; wherein the undirected graph is segmented into a plurality of connected components after preprocessing, so that a connected component set is obtained; the graph reduction strategy comprises a short edge deletion strategy and a connected component deletion strategy; in combination with a PD-LS search strategy of an ant colony algorithm, a longer circle can be obtained by local disturbance each time within finite time by changing an end condition of a depth-first algorithm, and when the ant colony algorithm is used for searching, a population is guided to develop in a better direction by continuously strengthening pheromones of edges in the circle searched by the ant colony before. According to the scheme, the time overhead is reduced, the method does not depend on the characteristics of the graph, the universality is higher, and the longest circle can be quickly and accurately found.

The invention combines an integrated circuit, a complex network theory and machine learning, and provides a method for predicting circuit performance by using machine learning. The method includes the steps of generating a data set, obtaining an optimized machine learning model and predicting the circuit performance by using the machine learning model. The data set generation part comprises the following steps: firstly, performing layout and wiring on an initial circuit by utilizing an EDA tool to obtain circuit performance and a wiring layout after wiring, then performing complex network modeling on the wiring layout, and finally extracting corresponding complex network characteristic parameters through a complex network analysis tool. The step of obtaining the optimized machine learning model comprises the steps of dividing a data set into a training set and a test set, training the machine learning model by using the training set, evaluating the obtained machine learning model by using the test set, and carrying out model optimization. The step of predicting the circuit performance by using the machine learning model comprises the steps of extracting complex network characteristic parameters of a to-be-tested circuit, inputting the complex network characteristic parameters into the optimized machine learning model, and predicting the circuit performance.