158 results about "Weighted network" patented technology

The present invention discloses a weighted network community clustering method based on hybrid measure to dig the clustering relationship of nodes in a large complex network. The method comprises a step of introducing a new node intimacy definition for measuring the association intensity between the nodes in a directed weighted network, a step of carrying out weighted processing on the side of a directed / undirected network through newly defined node intimacy, and a step of providing a modular new definition based on the node intimacy and using the hybrid measure to carry out hierarchical community structure detection on the directed / undirected network. Compared with the traditional community structure detection method, node relation information which can be referred in the communication division is increased by the hybrid measure, the quality of the communication division is improved, and the scale of an ultra large community is reduced. At the same time, the method provides a unified analysis method for the communication division of undirected unweighted, directed unweighted, undirected weighted and directed unweighted networks.

Systems and methods which provide network congestion avoidance implementing a fairness scheme in which a cost function is used are shown. Transmission cost, for use in network congestion avoidance packet dropping decisions, may be determined from an amount of air time needed to transmit a data packet, a transmission data rate, a size of a packet, an amount of spectral energy or transmit power associated with transmission, etc. A packet dropping probability for a particular packet is preferably determined as a function of the current flow cost and average flow cost to provide fair allocation of network communications resources. Embodiments additionally implement buffer based packet dropping probability calculation in order to provide a robust network congestion avoidance technique. Cost based packet dropping probability calculation and buffer based packet dropping probability calculation implemented according to embodiments are adapted to accommodate quality of service applications.

One embodiment provides a mobile wireless network that includes a plurality of wireless nodes and a controller node which manages a weighted network graph for the plurality of wireless nodes. A local wireless node sends a route-request message associated with at least one destination node to the controller node, receives a path to the destination node, and routes a packet to the destination node based on the received path. The path is computed based on the weighted network graph. One embodiment provides a system for routing in a mobile wireless network that comprises a plurality of wireless nodes. The system receives a route-request message associated with at least one destination node from a source node, computes a path between the source node and the destination node based on a weighted network graph for the plurality of wireless nodes, and transmits the computed path to at least the destination node.

The invention discloses a method for using a most influential node in a social network to achieve efficient viral marketing. The method for using the most influential node in the social network to achieve efficient viral marketing sequentially includes the steps of user definition and resource definition, establishment of edge connection for relationships between users and resources, establishment of interest strength between every two users, definition of importance measure indexes of weighted network nodes in boundary viewing dimension, calculation of a sum of edge weights in an NN assembly in a weighted network, establishment of a marketing model and the like. According to the method for using the most influential node in the social network to achieve efficient viral marketing, attention records between the users and the resources in the real social network are used, user interests are found, user interest strength relational diagrams are constructed, common attention of the users is used, on this basis, a half-partial centrality method based on neighborhood information is adopted to determine influential nodes, the nodes serve as initial nodes, spread of marketing information is carried out on a user interest network, key factors which can well restore interpersonal interaction processes in real life and comprise the accumulation effect, the social reinforcement effect and friend relationship strength are introduced into a spreading model, and the method for using the most influential node in the social network to achieve efficient viral marketing has the advantages of being low in calculated amount and good in robustness.

The invention relates to a method for evaluating the importance of nodes in a communication network, and belongs to the technical field of node analysis in the network. The method for evaluating the importance of the nodes in the communication network comprises the following steps: (1) establishing a mathematical model of a weighted network according to the actual communication network, (2) respectively calculating basic indexes, including a node degree k, a node betweenness b, a feature vector index Ce and a compactness index Cc, of the weighted network, conducting normalization, (3) conducting linear combined weighting on an F1, an F2, an F3 and an F4 to obtain a final score F of a comprehensive evaluation, (4) ranking the n nodes according to the value of the final score F of the comprehensive evaluation, using the nodes in the higher rank as the important nodes in the actual communication network, and therefore determining the importance of the nodes in the actual communication network. According to the method for evaluating the importance of the nodes in the communication network, firstly, the bandwidth is utilized for weighting the actual communication network, and then ranking of the importance of the nodes is achieved through the comprehensive evaluation.

The invention discloses a social network establishment method and a device, and a community discovery method and a device, and the social network establishment method comprises the following steps: respectively extracting feature words from all information units, and calculating feature vectors which correspond to all the information units according to the feature words; respectively calculating the similarity between each two information units according to the feature vectors; and establishing a social network according to the calculated similarity between each two information units. The method and the device can more really reflect the links among nodes in the network, and better carry out community division on the weighted network.

The invention discloses a multimode public transportation transferring method in an urban congestion period, which comprises the following steps of: step 1, constructing an urban ground-level road network based on real-time traffic information, and dividing different road sections of the urban road into a congestion road section and un-congestion road section; step 2, setting functional weight parameters for calculating costs of a ground bus, a subway and a public bicycle; step 3, establishing an urban public transportation weighted directed transferring network T; step 4, establishing an urban subway directed weighted network S; step 5, establishing a weighted directed public bicycle network B with the urban bus connected with the subway; step 6, by combining with the T network, the S network and the B network, calculating a cost function value, and adopting a breadth-first algorithm to obtain an optimal transferring solution. The multimode public transportation transferring method in the urban congestion period gives consideration to the characteristic that the short-distance tripping way by using the public bicycle and subway is not affected by the congestion of the ground road, accordingly provides the multimode transferring method to adjust the transferring solution automatically based on the congestion condition of the road surface, and enables the influence from the congestion to be minimum.

The invention discloses a weighting network node importance assessment method based on network heterogeneity. The method is based on the contribution entropy of nodes, and takes the network heterogeneity and the range of change of the network heterogeneity as an index for weighing the importance of the network nodes. The heterogeneity reflects the uniformity of a network. The more the heterogeneity of the network changes after some node is removed, the more important the node is on the network. Due to the fact that the network heterogeneity index which can reflect the overall macrostate is adopted, the limitations of focusing on details of the nodes with the existing public method in general use are overcome. The method not only can assess the importance of all the nodes accurately, but also solves the problem that misjudgment may be caused because cut points exist.

The invention relates to a mixed collaborative recommendation algorithm based on WNBI (weighted network-based inference with z-score normalization) and RSVD (regularized singular value decomposition). According to the method, a WNBI algorithm is firstly used for abstracting users and items into nodes in a network; information hidden in the network is used; deeper potential information between the items is mined; a neighbor set similar to the items is found; secondly, an RSVD model is used for decomposing a user-item grading matrix into a user feature matrix and an item feature matrix; the data density is improved through dimension reduction; finally, the item neighbor information of the items is used for regularizing the RSVD model so as to overcome the defects of a conventional method. The mixed collaborative recommendation based on WNBI and RSVD (RSVD_WNBI) can use the information, obtained by the WNBI algoritm, hidden in the user-item network for regularizing the RSVD model, so as to improve the recommendation accuracy and effectively solve the grading matrix sparsity problem.

The invention provides a method for network reconstruction double-layer optimization based on a node importance evaluation matrix and relates to a power supply network reconstruction method. At present, the weighted network node importance assessment result is too one-sided. The method comprises the steps of inputting an initial parameter of a particle swarm optimization; according to each particle, calling a double layer optimization model, wherein solving an upper layer optimization model to obtain a starting moment of a machine set and working out an available generating capacity of a system, solving a lower layer model to obtain a recovery path of a generator node, and thus obtaining a target function value of each particle; calculating the fitness of each particle according to the target function value; updating locations and speeds of the particles to obtain new particles; repeating the steps until the particle swarm reproductive generation number Mc is reached; selecting optimal particles, causing the solution to the upper layer optimization model corresponding to the optimal particles to be the optimal machine set starting time, and causing the solution to the lower layer model to be the recovery path. According to the technical scheme, the assessment of the node importance is more comprehensive, and the problem that the machine set delays the recovery is solved effectively.

The method is suitable for the technical field of computers, and provides a time link prediction method for a dynamic weighted network. The method comprises the following steps of obtaining a networktopological graph of the dynamic weighted network under the continuous historical time stamps; inputting the network topological graph under the historical timestamp into a pre-trained generative adversarial network; predicting to obtain a network topological graph of the dynamic weighted network under the to-be-predicted timestamp, wherein the generative adversarial network comprises a generativemodel and an adversarial model, and the generative model comprises an attention map convolutional network and an enhanced attention long-short-term memory network fusing the time matrix decomposition, an attention mechanism and a long-short-term memory network, thereby realizing the time link prediction of the dynamic weighted network, and improving the time link prediction accuracy and effect ofthe dynamic weighted network.

The invention relates to a method for assessing the importance of nodes in a power communication network based on fast density clustering, which belongs to the field of power communication systems. The method includes the following steps: establishing a data model of a power network based on the power communication network to be evaluated; counting network bandwidth and distance , normalize and integrate the weights of distance and bandwidth, and use network bandwidth and distance as evaluation weights; calculate node degree, calculate node closeness, calculate node betweenness, and normalize the calculated data; normalized The data is input into the fast density clustering algorithm, and the results of the importance of the nodes of the power communication network are obtained through analysis and calculation. Compared with the existing evaluation methods of node importance in power communication network, which often only use a single evaluation index, or use unsupervised classification of comprehensive index evaluation, there are many deficiencies. The fast-clustering-based evaluation method of the node importance of the power communication network in the present invention can effectively and quickly determine the node importance of the power communication network.

The invention relates to an analysis method of dynamic characteristics of electrocardiosignal, which comprises the following steps of: firstly, utilizing a surrogate data algorithm to carry out the dynamic characteristic recognition of collected signals; secondly, converting the recognized electrocardiosignal into a weighted network, and further capturing the difference between the dynamic characteristics of different types of electrocardiosignal in a framework of a complex network by a point intensity distribution map; and finally, defining a statistic Rs, and successfully distinguishing normal electrocardiosignal from atrial fibrillation electrocardiosignal based on the statistic. In the technical scheme provided by the invention, the classification accuracy of the electrocardisignal isimproved, and the information contained in the electrocardisignal is deeply discovered.

The present invention relates to a microblog big data interest community analysis optimization method based on a user experience. The method comprises a step of carrying out weighted reconfiguration of an original microblog network, a step of completing the community division of the reconfigured weighted network based on the discovery algorithm of a link community, and a step of using a hierarchical clustering algorithm to continuously merge two communities with a largest similarity, finally forming a link community through division, and generating a tree-shaped hierarchical diagram. Starting from aspects of interest modeling and community discovery, through analysing microblog content and a user behavior, a user is helped to find interested users and topics of the user. Compared with a traditional method, the accuracy, recall ratio and F value of the method of the invention are improved significantly.

The invention relates to a method for calculating a maximum power supply capability of an active power distribution network based on second-order cone relaxation, belonging to the field of power system optimized evaluation. The method related by the invention comprises the steps of: considering all scenes of different main transformer faults and introducing new parameters for expressing different fault scenes; establishing a maximum power supply capability calculation model of the active power distribution network based on the second-order cone relaxation, wherein an objective function of the model is to maximize a total load capacity in the active power distribution network subtracting a weighted network power loss and simultaneously needs to satisfy each technical constraint for power network operation; and setting a second-order cone constraint for describing a load flow variable relationship in the active power distribution network. By solving the model, an evaluation result on the maximum power supply capability of the active power distribution network can be obtained, namely, a total load upper limit under the condition that distribution transformer N-1 constraints are met, and a load value of each node correspondingly when the load upper limit is used. The evaluation model established with the method of the invention is very precise, suitable for practical situation and easy to solve, and has very strong practicability.

The invention relates to a community discovery method combining node information and a network structure. The method comprises the following specific steps of (1) classifying node characteristics according to the influence of the node characteristics on community division; (2) carrying out content similarity calculation on nodes according to the node characteristics; (3) obtaining an adjacent matrix A of a network according to the network structure; (4) setting a threshold, updating a network weight and generating a weighted network; and (5) setting parameters according to the actual needs, selecting a community discovery algorithm and processing the weighted network obtained in the step (4) to obtain final community division. Through a matrix sum form, the node characteristics and the network structure are fused to a weight form, an unweighted and undirected network is transformed to a weighted and directed network; in addition, through a way of setting the threshold, the unnecessary computation cost is reduced and the time of a community discovery process is saved.

The invention discloses an implementation method for discovering important users of a weighting network. When solving the ability of different users in a network to influence each other, the method introduces second-layer neighbor node topology of a node to define expenses for maintaining an edge relation of the node, and the method comprehensively considers node bridging importance and node centrality, thereby obtaining importance of the node. The method adopts a structure hole theory to measure the importance of the node, fully considers the centrality of the node, introduces the second-layer neighbor topology of the node, and distinguishes importance of bridging points of different properties. The method comprehensively considers bridging and centrality functions of the node, thereby improving accuracy of an algorithm.

The invention relates to a brain function connection module division method based on the weighted network. The method mainly comprises steps that brain function magnetic resonance imaging pre-processing and partition template matching standardizing are carried out, and the time sequence corresponding to each brain zone is extracted; the sub time sequence corresponding to each window is separated through the sliding window method, correlation coefficient matrixes of all the windows are combined, and the dynamic weighted network of brain function connection is constructed; the edge betweenness of the weighted network during weight considering is acquired through the betweenness rate, the ratio of the edge betweenness ignoring weight, the betweenness rate and the connection edge weight is calculated, and the connection edge with the highest ratio is removed; the module division result is outputted and a modularity value is calculated till no connection edge in the network can be removed;the module division result corresponding to the largest modularity value is outputted. The method is advantaged in that dynamic characteristics and weight change of brain function connection are comprehensively considered, and shortcomings of the traditional module division method of ignoring time-varying characteristics and simple thresholding of the weight are made up.

The invention relates to a cyberspace security situation awareness analysis method and system. Sufficient mining analysis of a security situation time sequence is carried out; a directed weighted network representing a security situation state change rule is constructed; the directed weighted network is analyzed, so that attribute data is obtained; therefore, the security situation change rule contained in cyberspace security situation data is analyzed effectively; the problem that analysis on existing data statistical modelling currently is limited can be made up; the network security situation change rule is sufficiently mined; furthermore, the network security situation can be reflected intuitively; support is provided for high-speed decision of network managers; and thus, the network can be relatively safe and effective.

The invention provides an adjacent matrix based dependent network system and a frangibility detection method thereof. Network structure information of the dependent network system is expressed by an adjacent matrix, and the dependent network system comprises a first network which is a weighted network and a second network which is an un-weighted network; in each of the first and second networks, nodes are connected by connecting sides to express connection relation between the internal nodes; and between the first and second networks, cross-network nodes are connected by dependent sides to express the dependent relation between the network nodes. Thus, the frangibility of the dependent network system can be detected effectively.

The invention discloses a multivariable clustering and fusion time series combination prediction method; aiming at solving the problems that an existing neural network model does not have a specific learning mechanism and cannot fully mine data structure feature information, from the multivariable directed coupling perspective, and in combination with the advantages of a graph convolutional neuralnetwork and a long-term and short-term memory network, the invention provides a multivariable clustering and fusion time series combination prediction method. The method comprises the following steps: firstly, exploring a causal transfer relationship between variables based on coupled Granger causal measure analysis; secondly, establishing a directed weighted network according to a variable causality analysis result, extracting node and edge weight characteristics of the directed weighted network, and embedding the weight of a target variable into a graph convolutional neural network for training to realize accurate classification of monitoring variables; finally, taking the non-target monitoring variable time series contained in the community where the target monitoring variable is located as input, and predicting the target monitoring variable based on the long-term and short-term memory neural network. The method is applied to verification of a compressor unit monitoring sequence in a chemical production system, and results show that the method is superior to a traditional node classification method in the aspects of prediction accuracy and calculation complexity, and the proposed method can also maintain high prediction capability in an abnormal state of the system.

The present invention relates to a parallelization critical node discovery method for postal delivery data. The method comprises the following steps: step S1: acquiring node activity according to the total number of sending and receiving times of each node in a set time in the postal delivery data, and taking the node activity as the own weight value of the node; step S2: acquiring the weight values of edges of each node pair according to the interaction frequency and shared neighbor number metric indexes of each node pair in the set time in the postal delivery data, and defining a network formed by the postal delivery data as a directed double-weighted network graph; and step S3: adding the own weight values of the nodes and the weight values of the edges of the node pairs on the basis of a PageRank algorithm, and excavating critical nodes in the directed double-weighted network graph in parallel. In contrast to the prior art, the parallelization critical node discovery method fully utilizes information in a logistics postal delivery network, reduces the loss of useful information, improves the accuracy of discovery of critical nodes in the network, and parallel operation is implemented at the same time, thereby greatly improving the efficiency and stability of critical node excavation.

The invention discloses a sequential network node importance mining method and device; the method firstly builds the sequential network according to interaction time relations between N nodes, and then slices the sequential network according to preset time window sizes, thus obtaining sliced network sequence; the sliced network sequence comprises the following steps: for m slice networks matched with m time windows, a weight value is respectively calculated for all connecting edges of each slice network; weight value accumulation is respectively carried out according to different connecting edges for weight values corresponding to all connecting edges in each slice network, thus obtaining the weight accumulated value of the connecting edge between paired nodes in the sequential network; the weight accumulated value of the connecting edge between each paired nodes is updated to the sequential network, thus obtaining the weighted network; finally using the connecting edge containing the weight accumulated value to calculate importance indexes of each node in the weighted network, wherein the importance indexes of each node comprise the sum of weight accumulated values corresponding to all connecting edges of the corresponding nodes.

The invention provides a neural network training method and device suitable for a long-tail distribution data set. The neural network comprises a feature extraction network, a classifier and a category gradient reweighting network. The training method comprises the following steps: obtaining a training sample set; extracting, bythe feature extraction network, features from the training sample setto obtain features, classifying the features through a classifier, and establishing a loss function according to a classification result; calculating the gradient of each neuron in the feature extraction network in the training sample according to the loss function; and in the back propagation process of neural network training, calculating, by the class gradient reweighting network, the reweighting gradient weight of the training sample, and adjusting the gradients of the training samples belonging to different classes according to the reweighting gradient weight. Therefore, the method solvesa problem that the recognition accuracy of the neural network is reduced under the training data of the long-tail distribution, alleviates the overfitting phenomenon of the feature extraction network, and improves the recognition accuracy and robustness of the deep neural network under the long-tail distribution.

The invention discloses a community self-organizing detection method for power network fault diagnosis. The method comprises the steps of firstly, collecting network characteristic parameters of power networks, then describing the power networks as weighted network models, constructing local fitness and global fitness functions, starting from grouped solutions of the power networks, which are generated randomly, calculating local fitness of each power node, sequencing the local fitness, selecting the nodes with the poor local fitness according to an expansion evolution probability distribution function, transferring the nodes with the poor local fitness to another group of networks to generate new solutions, comparing global fitness values of the new solutions and the current solutions, reserving the best solutions in the new solutions and the current solutions, enabling the new solutions to serve as initial solutions for the next iteration to repeat above optimization processes until preset end conditions are met, and finally, analyzing and outputting community self-organizing detection results which are used for power network fault diagnosis. Compared with conventional methods, the method has the advantages of being a few in adjusting parameter, simple in detection process, easy to implement and high in detection efficiency and detection precision.

The invention provides a software network key node mining method based on a complex network. The method comprises the steps of firstly according to a network topological structure, taking the number of the methods in a class as the weights of the edges, and redefining the concept of the edge weight value, namely weighting the edges in a directed network of a software system according to the numberof the methods of the class to abstract a directed weighted network model; and then using the nodes which are judged as the alternative key nodes by different key node mining algorithms as the finalkey nodes in the directed weighted network model, so that the obtained key nodes are the key nodes occupying more important positions in the software network, and by protecting the found key nodes ofthe software system, the reliability and the safety of the software system can be enhanced, the attacks from the outside on the software system are effectively reduced, and the loss caused by the damage to the system is greatly reduced.

The invention relates to a network reconfiguration method based on a geographical position. The network reconfiguration method comprises the following steps: step S1, calculating the node network structure similarity; step S2, calculating the geographical position similarity of a user; step S3, combining the node network structure similarity and the geographical position similarity of the user, establishing the uniform similarity; and step S4, adopting a threshold processing method to filter the uniform similarity obtained in step S3, and reconstructing a weighted network according to a filter result. The network reconfiguration method based on the geographical position provided by the invention reconstructs a social network by combining with a dynamic characteristic of the geographical position of the user, and a reconstructed network graph is easier to obtain the network characteristics with position features; a result of community discovery based on a weighted network structure has community division for geographical position information; and the community discovery is performed after construction, and communities which are distributed on the geographical position in a relatively concentrated way can be obtained.

The invention discloses a PPI (Point-Point Interaction) network clustering method based on an artificial swarm reproduction mechanism, comprising specific steps of: converting a PPI network into an undirected weighted graph; setting parameters; pre-treating each knot and each edge of the PPI network; calculating a weighted network comprehensive characteristic value of all the knots; initializing queen bees; carrying out a mating flight process; partially searching young bees; optimally selecting the queen bees; and selecting the current fitness and comparing until a global optimum clustering result is output. According to the method disclosed by the invention, the clustering quantity does not be pre-set and can be automatically obtained in a clustering process, so that the subjectivity of artificially setting the clustering quantity is avoided, and the time complexity is obviously reduced. An MIPS (Million Instructions Per Second) database is used for carrying out experiment simulation, a result is closer to a standard database, and indexes including the accuracy, the recall ratio, the operation time and the like are better. Compared with the other clustering methods, the method can automatically determine the clustering quantity by adopting the artificial swarm reproduction mechanism based on the reproduction mechanism, so that the clustering process is realized, and the clustering effect and the calculation efficiency are effectively improved.

The invention belongs to the technical field of network information processing. The invention discloses a microblog social circle mining method and system based on an artificial immune network, and the method comprises the steps: carrying out the calculation of the similarity between users through the analysis of social information and interest information between the users, measuring the relationship strength between the users through the similarity, and carrying out the comprehensive description of the relationship strength between the users; And on the basis, constructing a microblog undirected weighted network taking the users as nodes and the relationship strength as weights, and removing edges with relatively low relationship strength in the undirected weighted network to obtain a microblog user similarity network. According to the invention, an artificial immunization method with relatively high adaptability and self-adjustability is adopted; A principle and an action mechanismof a biological immune network are applied to similarity clustering of microblog users, users with high relationship strength are divided together, and one user is allowed to belong to a plurality ofsocial circles at the same time during division, so that mining of overlapped social circles is realized.