407 results about "Graph theory" patented technology

The invention, referred to herein as PeaCoCk, uses a unique blend of technologies from statistics, information theory, and graph theory to quantify and discover patterns in relationships between entities, such as products and customers, as evidenced by purchase behavior. In contrast to traditional purchase-frequency based market basket analysis techniques, such as association rules which mostly generate obvious and spurious associations, PeaCoCk employs information-theoretic notions of consistency and similarity, which allows robust statistical analysis of the true, statistically significant, and logical associations between products. Therefore, PeaCoCk lends itself to reliable, robust predictive analytics based on purchase-behavior.

The invention, referred to herein as PeaCoCk, uses a unique blend of technologies from statistics, information theory, and graph theory to quantify and discover patterns in relationships between entities, such as products and customers, as evidenced by purchase behavior. In contrast to traditional purchase-frequency based market basket analysis techniques, such as association rules which mostly generate obvious and spurious associations, PeaCoCk employs information-theoretic notions of consistency and similarity, which allows robust statistical analysis of the true, statistically significant, and logical associations between products. Therefore, PeaCoCk lends itself to reliable, robust predictive analytics based on purchase-behavior.

Embodiments of a system and method for providing class dependency graph-based class loading and reloading may be used to segregate namespaces in a graph-centric way, and may provide a set of normalized topologies that may be used to efficiently support hot-swapping of programmatic logic such as classes, applets, and beans, among other applications. Embodiments may provide a domain-independent, flexible and robust namespace segregation technique that is based on the dependency between the various classes and not on details like the roles the classes play. The problem of segregating namespaces is formulated as a graph theory problem, and a solution is sought through graph techniques. The graph may be normalized by identifying and grouping interdependent classes and non-interdependent classes in separate groups. A directed dependency relationship of the groups may be determined using the relationships between the member classes of the groups.

The invention provides a multi-camera system calibrating method based on an optical imaging test head and a visual graph structure. The method comprises the following steps: independently calibrating each camera by the optical imaging test head to obtain the initial values of the internal parameter and aberration parameter of each camera; calibrating the multiple cameras two by two, and obtaining the fundamental matrix, polar constraint, rotation matrix and translation vector between every two cameras with a plurality of overlapped regions at a view field by means of linear estimation; building the connection relationship among the multiple cameras according to the graph theory and the visual graph structure, and estimating the rotation vector quantity initial value and translation vector quantity initial value of each camera relative to the referred cameras by a shortest path method; and optimally estimating all the internal parameters and external parameters of the all cameras and the acquired three-dimensional sign point set of the optical imaging test head by a sparse bundling and adjusting algorithm to obtain a high-precision calibrating result. The multi-camera system calibrating method is simple in a calibrating process from the partial situation to the overall situation and from the robust to the precise, ensures high-precise and robust calibration, and is applied to calibrating multi-camera systems with different measurement ranges and different distribution structures.

The invention discloses a method for recognizing a human motion in the field of computer vision and pattern recognition. Firstly, a feature containing time and space information is used for expressing the motion status of a human body in a current frame, and a classifier is designed through a graph theory semi-supervised method so that the purpose of human motion recognition is achieved. In the process of extracting the human motion feature, outline and motion light stream information in the past, in the present time and in the future is simultaneously fused so that the motion posture of a human body can be more accurately described. In addition, in order to obtain a high recognition rate with a few samples, based on the graph theory semi-supervised method of the generalized Laplacian matrix, and the graph theory semi-supervised method is used for recognizing the human motion. Experiments prove that under the conditions that observation angles are different and the motion difference of different persons exists, the method can be used for obtaining the satisfactory recognition rate of common motions.

A data infrastructure for graph-based computing that combines the natural language expressiveness of the Semantic Web and the mathematical rigor of graph theory to discover meaningful associations across multiple sources towards computer-assisted serendipitous insight discovery. The process automatically integrates massive size datasets accessed using Semantic Web standards and technologies and normalizes data in graphs. The process generates a plurality of conditional probability distributions based on type-triple meta-data and triple statistics to model saliency and automatically construct and evaluate a plurality of sub-graphs based on the plurality of conditional probabilities for contextual-saliency. The process then renders a plurality of paths (i.e. sequence of associations) that model meaningful pairwise relations between objects of the normalized integrated data. The pluralities of conditional probabilities reveal and rank previously unknown associations between entities of user-interest in the knowledge graph.

A system and method for reconstructing pathways in large genetic networks from genetic perturbations comprises an analysis method and system that applies a recursive algorithm for determining the path between every gene pair in an arbitrarily large genetic network from large-scale gene perturbation data and reconstructs all direct and indirect regulatory gene interactions in the network. Graph theory mathematics is applied to genetic network reconstruction in the following manner: Genetic perturbation data is used to identify all genes accessible from a perturbed gene to generate an accessibility list for the gene. Graph theory mathematics is applied to the accessibility list and its graph to determine a condensation of the graph as defined by the condensation's accessibility list. Graph theory mathematics is applied to the accessibility list, such as through a recursive algorithm performed on a desktop computer, to obtain an adjacency list for the gene that characterizes a genetic network.

The invention discloses an SDN multi-controller deployment novel method for reducing management load overhead, belonging to the technical field of the software defined network (SDN). The method is characterized in that the management load of a software defined network (SDN) controller is used as a decision variable, and a mathematical model about controller management load is constructed. An original SDN network multi-controller deployment problem is abstracted to be a graph theory problem, the graph theory problem is converted into an integer linear programming problem through establishing the mathematical model, and an NP difficult problem is solved by using an approximate algorithm. According to the method, in the process of constructing the mathematical model of the multi-controller deployment problem, the network performance and the management load of the SDN controller are creatively taken into consideration, effective deployment schemes can be provided for different SDN networks, the reasonable selection and effective deployment of the number of multiple controllers are realized, and the normal and effective operation of the SDN network in the condition of the satisfaction of certain network performance and minimum cost are ensured.

The invention discloses a method for predicting dynamic social network user behaviors based on a computer probability graph model, which comprises the following steps of: 1, performing objective statistical analysis on the dynamic social network user behaviors in terms of social influence, time dependence and network correlation; 2, performing formal definition on the dynamic social network user behaviors by adopting computer technical means such as a graph theory, a set, a matrix theory and the like; 3, establishing a dynamic anti-noise factor graph model according to the definition in the step 2; 4, learning the dynamic anti-noise factor graph model, and estimating a value Theta of a series parameter from given historic records; and 5, predicting the user behaviors according to the Theta to obtain prediction results. By the method, modeling and accurate prediction are performed on the dynamic social network user behaviors from a micro level.

The invention provides a method for classifying sports video based on a key frame of a main scene lens. The method only adopts the main scene to perform sports classification rather than classification by frames of the whole video to represent the sports video so as to effectively reduce the calculation amount of video classification. The method comprises the following steps that: firstly, the video is automatically divided into a plurality of fragments according to the lens, key frames of all the fragments are classified into a plurality of types including long shot, medium shot and close-up after subjected to adaptive threshold cluster based on graph theory, wherein the medium shot type is used as the main scene lens of the sports video, main scene information of the sports video, namely the medium shot, can be automatically and effectively extracted without depending on any prior information in the process; multiple interferences (such as judges, close-up of spectators, broadcasting effects, advertisement and other lens) in the sports video are removed; and finally, the key frame of the main scene lens is classified by an SVM classifier. The method has high accuracy for classifying the sports video.

The invention discloses a method for scheduling parallel test tasks based on grouping and tabu search, belonging to the technical field of automatic test and measurement. The method successively comprises the following steps: a restrain relation among test tasks is determined and analyzed; a graph theory model is built to group the test tasks; a peak dyeing theory in the graph theory is adopted to process the grouped test tasks; an initial scheduling sequence of parallel test task scheduling is configured and tested in parallel according to the group result; and iteration search is carried out by using a tabu search method to search an optimum scheduling sequence, early test starting time of each task is successively determined according to the optimum scheduling sequence, thereby completing a task scheduling plan based on the shortest test time. In the method, the indeterminacy of initial value selection in the process of searching in the traditional method is solved by the method through combining the practical problem of parallel test task scheduling and analyzing the characteristics of the tabu search method, so that the initial scheduling sequence can be well configured and the search rate of the method can be improved, thereby fast finding the optimal task scheduling scheme.

The invention discloses a method for identifying protein functions based on a protein-protein interaction network and network topological structure features. Firstly, a node and side-weighted protein-protection interaction network is established, wherein the node represents protein while the edge represents the interaction; then the nodes and the sides in the network are weighted by protein first-grade structural description and protein-protein interaction trust scoring; protection functional annotation data is collected to establish a data set, and a new protein with overall and local information network topological structure features is provided based on a graph theory; and finally, the protein functions are predicated by choosing features through adopting a minimum-redundancy maximum-correlation method and by modeling through a support vector machine. The protein function predication method is greatly better than the prior art, and has robustness on sequence similarity and sampling; and meanwhile, information of three-dimensional structure and the like of protein is not required, so that the method is simple, rapid, accurate and efficient, and the method is expected to be applied in the research fields of proteomics and the like.

The invention relates to a method for automatically acquiring key sections based on electrical partitioning, which belongs to the technical field of operation and control of electric systems. The method comprises the following steps: establishing a relation of connection among substations in a power grid by using topological research, and forming a network N by taking the substations as vertexes and taking transmission lines as edges; generating an impedance matrix Z of the network N by using a branch-adding method, and obtaining the electric distances among the substations; partitioning the network N into multiple electrical partitions and adjusting the partitions according to the geographical distribution of the substations in the network N; obtaining initial sections according to a graph theory; screening the initial sections and than obtaining transmission sections; and judging the safety margins of the transmission sections so as to obtain the key sections. The method of the invention can better adapt to the increasingly volatile operation modes of the power grid, provides more accurate transmission sections for online fine-rule-making, and improves the fineness degrees and online adaptability of fine rules.

The invention relates to a pedestrian intention recognition method based on graph convolution. The video image of road environment is captured by a forward-looking camera system mounted on a vehicle.Pedestrian detection and pedestrian human key points extraction are carried out on the image, and the adjacency matrix is constructed to represent the connection information of pedestrian human key points based on graph theory. The bottom feature is extracted from coordinate information and adjacency matrix representation of key points by graph convolution algorithm, and the bottom feature is extracted and time series analyzed by depth convolution neural network and depth loop neural network. Based on the data set of pedestrian intention constructed by manual labeling method, the parameters ofthe model are optimized to realize the classification and recognition of pedestrian intention. The invention effectively utilizes the high-level semantic feature of the key point information of the pedestrian human body, so that the automobile advanced driving assistance system has the ability to understand the pedestrian behavior intention.

The invention discloses an intelligent carrying robot optimal path hybrid graph theory control planning method. According to the method, the idea of hierarchical planning for the global path is put forward, local optimal path planning of each room or passageway acts as the minimum planning unit to search the optimal hub node of each floor, the local optimal paths of all the rooms or the passageways are connected to obtain the optimal path of a single floor, and finally the optimal paths between the floors are connected so as to obtain the complete global optimal path. The computational burden in the path planning process can be greatly reduced by the design of hierarchical planning, and path planning can be rapidly realized.

This invention belongs to intelligent font and graph information process technique and in detail relates to a method of paper page headline and cross logic connection, which comprises the following steps: first to establish a mathematics model with graph theory; to use bisect graph matching model to prescribe non cross area and cross area matching particle with one to one characteristics; to establish the weigh bisect graph according to space relationship; to firstly adopt nature language process technique to compute the bisect graph weigh value; to make the optimized result pair saturation top as logic connection success headline and content page.

The invention pertains to a spectrum resource management technique in cognitive radio networks, which is an OFDM spectrum allocation method for cognitive radio networks on the basis of receiver interference, comprising the steps of initialization, coloring primary users, carrying out topology updating, calculating tag values and colors to be put on of each secondary user, putting the color on the secondary user with the maximum tag value, etc. The invention discloses an OFDM spectrum allocation method for cognitive radio networks under a receiver interference model on the basis of graph theory, which can accurately model the spectrum allocation for cognitive radio networks and carry out spectrum allocation effectively in cognitive radio networks, and the target of the method of the invention can maximize spectrum efficiency and maximize minimum fairness or maximize proportional fairness. The method disclosed can be realized by in a distributed manner.

The invention requests for protecting a resource distribution method based on full-duplex D2D communication in a cellular network. Full-duplex (FD) D2D (Device-to-device) communication is introduced in the cellular network; therefore, the burden of a base station can be reduced; the energy consumption of a terminal can be reduced; the cell coverage range can be enlarged; but, resource sharing of D2D users and cellular users brings about a certain same-frequency interference to a system; mutual interference can be reduced through reasonable resource distribution; a resource distribution algorithm having the maximum weight matching is provided in the invention; on the premise that the service quality requirements of the cellular users are ensured through the algorithm, the lowest requirements of D2D link communication can also be ensured through the algorithm; the resource distribution relationship of the cellular users and the D2D users in the system is equivalent to the vertex matching relationship in a graph theory; wireless resources are reasonably distributed between D2D user pairs and the cellular users by utilizing the maximum weight matching algorithm in the graph theory; interference between the cellular users and the D2D users can be effectively reduced; the relatively good fairness can be obtained; and simultaneously, the throughput capacity of the system is maximized.

The invention discloses a power grid differentiation-based core backbone network architecture construction method. The importance of lines and nodes in a power system are taken into comprehensive account, the knowledge of a risk theory is referenced, the importance of the lines is obtained by calculating risk indicators of the lines, and the importance of the nodes is evaluated based on the knowledge of a graph theory; and a core backbone network architecture searching model is constructed based on the importance of the lines and the nodes, and network architectures are searched by a particle swarm algorithm to obtain the lines and the nodes for constructing a core backbone network architecture. The importance of the lines is evaluated by line returning risks, so that the importance of the lines can be better reflected from the point of electrical quantity variation; and a fuzzy-membership-based line returning risk evaluation method more consistent with actual conditions is provided, and the influence of uncertain factors of the power system is taken into comprehensive account.

The invention discloses a method for finding an optimal traffic route under disastrous environments. The method includes: establishing a road traffic network topological graph taking cities as nodes and roads connecting the cities as arcs according to the graph theory and actual traffic geographic information, wherein the actual traffic geographic information includes length, degree of reliability and capability of intercity roads; when a certain area in the road traffic network topological graph is in a disaster, calculating disaster situation of each road in the road traffic network topological graph according to area and strength of the disaster, and updating the length, the degree of reliability and the capacity of the roads according the disaster situations; utilizing the network algorithm to extract m feasible routes from a node O to a target node D from the disaster area; and choosing the optimal route from the m routes according to the updated length, degree of reliability and capacity of the roads. By adopting the method, an effective auxiliary decision-making basis can be provided for evacuation guiding and rescue directing under the disastrous environments.

The invention discloses an entity attribute information extraction method and device based on syntactic dependency. The method comprises the steps that firstly, a to-be-extracted text is preprocessedto obtain a to-be-extracted text entity; then, according to the syntactic dependency and the part-of-speech relation of the to-be-extracted text, an undirected weighted graph between words is established, and candidate attribute information of the to-be-extracted text entity is obtained according to the part-of-speech relation; the shortest path between the to-be-extracted text entity and the words of the candidate attribute information is searched for, and the words passing through the shortest path form an association information word set; finally, the semantic similarity between each attribute in the attribute set and the association information word set is calculated, an entity attribute is obtained, and the entity, the entity attribute and the attribute information are integrated to serve as a final extraction result. The natural language processing technology and the graph theory model are combined, the ambiguity of text information is solved, and the text extraction accuracy isimproved; the semantic similarity of the keywords is utilized, the attributes of the abstract information are automatically summarized, and the extraction efficiency is improved.

The invention discloses an AGV (automated guided vehicle) scheduling method and system adopting inertial navigation based on a complex path, and belongs to the field of automatic control. The method comprises the steps as follows: determining an optimal distribution task among a plurality of distribution tasks according to obtained delivery task information with a weighting algorithm; determininga first path, on which the distance between an AGV and a starting node position is the shortest, and a second path, on which the distance between the starting node position and a target node is the shortest, with a graph theory process according to the current node position of the AGV as well as the starting node position, the target node position and a transport network model corresponding to theoptimal distribution task; according to the first path, the second path and a priority-based traffic rule, driving the AGV from the current node to the starting node position, and driving the AGV from the starting node position to the target node position. The method provides an effective, complete and feasible solution for task scheduling and path planning of an AGV system.

The invention discloses a deadlock detection method and a device of a database transaction lock mechanism, which is characterized by predefining an adjacency matrix for storing cross-thread waiting relation messages. The method comprises the following steps: a locking thread records the cross-thread waiting relation messages generated in the process of locking in the adjacency matrix; an unlocking thread updates corresponding waiting relation messages in the adjacency matrix in the process of unlocking according to the requirements; a deadlock detection thread detects and calculates the thread by adopting the adjacency matrix and principles of graph theory so as to judge whether deadlock exists. The device comprises a message storage module, a deadlock detection module, and a message record module and a message update module. The technical proposal of the invention has extremely high deadlock detection speed, can fully utilize useful messages obtained in the processes of locking and unlocking to assist subsequent deadlock detection and saves calculation resources.

The invention discloses an intelligent control method for earth pressure balance shield machine tunnel piercing parameters. The intelligent control method is characterized by comprising the following steps of (S1) collecting data and extracting factor nodes affecting shield machine piercing performance, (S2) constructing a topology structure for the factor nodes affecting the shield machine piercing performance to acquire a fuzzy association matrix Wij of each node, (S3) constructing a fuzzy control network map of the earth pressure balance shield mechanism piercing, and dynamically evolving and calculating a state value of each factor node at a t+1 time of period via an iteration rational formula, and (S4) conducting multi-phase decision analysis according to the fuzzy control network map to real-time analyze and control deviation rectification of the earth pressure balance shield machine tunnel piercing parameters. Expert prior knowledge, fuzzy logic and graph theory technologies are employed and ratiocination and real-time analysis and decision make are provided for multi-phase shield machine parameters before, amid and after subway shield machine engineering construction under complicate environment.