142 results about "Time clustering" patented technology

A system for providing real-time cluster configuration data within a clustered computer network including a plurality of clusters, including a primary node in each cluster wherein the primary node includes a primary repository manager, a secondary node in each cluster wherein the secondary node includes a secondary repository manager, and wherein the secondary repository manager cooperates with the primary repository manager to maintain information at the secondary node consistent with information maintained at the primary node.

The present disclosure is directed towards systems and methods for efficiently identifying periodic trends of user-defined annotations among users of a social networking system. For example, systems and methods described herein involve grouping a set of digital content items into subsets of digital content items and performing one or more iterations of a clustering algorithm on the grouped set of digital content items to identify one or more temporal clusters of the user-defined annotation. Additionally, the systems and methods described herein involve performing one or more additional iterations of the clustering algorithm on one or more reduced sets of digital content items to identify one or more additional temporal clusters of the user-defined annotation. Further, the systems and methods involve determining that the user-defined annotation is periodic based on a time period between identified temporal clusters.

The invention discloses a Trojan horse detection method based on communication behavior clustering, and belongs to the field of information safety. The unknown Trojan horse detection method is excellent in feature extraction performance, proper in clustering algorithm and high in detection efficiency and accuracy in order to resolve the problems that the existing Trojan horse detection technology is low in feature extraction capacity, improper in clustering algorithm selection and the like. According to the technical scheme, the Trojan horse detection method comprises the steps of extracting a network flow data package, recombining a TCP conversation, extracting a Trojan horse reverse connecting feature, an entropy feature, a heart beat feature and the like, building a feature vector of the TCP conversation and carrying out real-time clustering on the feature vector based on a real-time increment clustering algorithm of LSH. According to the difference of communication behavior features of a Trojan horse conversation and normal network communication behaviors, the Trojan horse detection method marks the difference of the communication behavior features of the Trojan horse conversation and the normal network communication behaviors by combining the statistic analysis and the time series analysis technology, guarantees high detection accuracy and a zero false alarm rate, lowers the false alarm rate, and can effectively carry out real-time detection on the abnormal communication behaviors of a Trojan horse.

The invention relates to a real-time land cluster survey method. The method comprises the following steps: with an office data service platform as a center, firstly sending the filed data acquired by virtue of GPS / PDA (global positioning system / personal digital assistant) and GPS (global positioning system) data used for monitoring and positioning in a field data acquisition subsystem to a real-time communication service system in a standard data definition format by virtue of the TCP / IP (transmission control protocol / internet protocol) network protocol; classifying the obtained data by the office data server; and issuing spatial data and attribute data in the whole database by a service terminal, and browsing and analyzing to generate a report for use of a user, thus the real-time cluster survey monitoring on the land is realized finally.

The invention discloses an indoor positioning method based on reception signal intensity clustering and reference point position clustering, solving the problem that the reference point clustering of the current indoor positioning method is not accurate and positioning accuracy is poor. The indoor positioning method disclosed by the invention comprises steps of (1) choosing a reference point, measuring reception signal intensity and storing into a database, (2) performing first time clustering on all reference point according to the reception signal intensity, (3) performing second time clustering on the cluster obtained through the first time clustering, (4) calculating the distance between the clusters obtained through the second time clustering, wherein the clusters having small distances are combined into one cluster, (5) measuring the reception signal intensity of a point to be positioned and roughly positioning the matched clusters, and (6) utilizing the roughly positioned matched clusters and sensing the precise positioning through compression. The invention reduces the positioning errors, improves the positioning accuracy and is used for the indoor positioning of the Wi-Fi reception machine.

The embodiment of the invention discloses a visitor flow rate statistical method and device based on image recognition, and the method comprises the steps: sequentially obtaining target human faces in N target images, so as to obtain a first target human face set, wherein N is a positive integer; carrying out the filtering of the first target human face set based on a human face tracking algorithm, so as to obtain a second target human face set; carrying out the filtering of the second target human face set based on a human face feature value real-time clustering algorithm, so as to obtain a third target human face set; and counting the target human faces in the third target human face set, so as to obtain a visitor flow rate statistical value. The method carries out the two-layer filtering of N target images, and can accurately filter out repeated human faces in the target human faces, so as to avoid the repeated counting of target human faces, and improve the visitor flow rate statistical accuracy.

Precision and reliability of localization estimates derived from conventional localization systems are improved through a system and method for post-processing of initial localization data, even in environments which may include noise, reflections, or other interference. Such localization systems include conventional sound source localization (SSL) systems based on microphone array inputs, radio source location systems based on directional antenna array inputs, etc. In general, this post-processing system and method applies statistical real-time clustering to initial localization estimates, and then uses this real-time clustering in a multi-stage process to generate new localization estimates having improved precision and reliability relative to the initial localization estimates.

System and methods for rapid generation of attribute driven temporal clustering are provided. In one embodiment, the system includes a data storage device and a server. The data storage device may be configured to store a database comprising a plurality of records. The system may also include a server in data communication with the data storage device. The server may be suitably programmed to receive a first attribute and a second attribute, search a database stored on a data storage device to obtain a first group of records associated with the first attribute, search the first group of records to obtain a second group of records associated with the second attribute, determine a temporal relationship between a first index date of a first attribute and a second index date of the second attribute, and generate an output comprising a graphical representation of the temporal relationship.

The invention discloses a voiceprint model updating method and system. The method comprises that present login time of a target speaker and last update time of a voiceprint model of the target speaker are obtained; a time period between the last update time of the voiceprint model of the target speaker and the present login time of the target speaker is divided into time clusters; voice data of the target speaker, which is obtained when voice authentication succeeds, in each time cluster is acquired; voice data is selected form voice data which is obtained when authentication succeeds in each time cluster and serves as update data for the voiceprint model of the target speaker; a voiceprint model is retrained by utilizing the update data for the voice model of the target speaker and training data of an original voiceprint model, and a trained new voiceprint model is obtained; and the new voiceprint model is used to update the voiceprint model of the target speaker. According to the invention, a pretender is prevented from updating the voice model of the target speaker continuously in a very short time, and it is ensured that the voiceprint model of the target speaker is updated correctly.

The invention discloses a method for determining relation between the built-up environment and rail passenger flow distribution based on signaling data. On the basis of mobile phone signaling data generated by individuals every day, a space-time clustering algorithm is used to sort the mobile phone signaling data of the rail station passenger flow according to the personal identification number and the triggering time; the data is preprocessed, identification of distribution places of passenger flow around a rail station point is completed by using a space-time threshold algorithm, and hot spot areas of the rail passenger flow distributed around the station are identified by using a space-time clustering algorithm rule and taking the number of rail passengers of a base station as a clustering cluster. Then, on the basis of extracting the number of various public infrastructures in the urban built-up environment in the hot spot area, a correlation structure equation model of the passenger flow distribution number and the number of built-up environment facilities is established, and the correlation relationship between passenger flow distribution and the built-up environment is analyzed; the method can be used for evaluating the influence degree and the correlation relationship of the built environments of different regions of the city on the rail passenger flow distribution.

The invention discloses a wireless sensor network data transmission method based on a double-layer fuzzy algorithm. The wireless sensor network data transmission method comprises the following steps:initializing a network and determining a node position; determining an energy consumption model; determining the quantity of cluster heads; carrying out clustering by adopting an improved algorithm; carrying out fuzzy knowledge processing to determine the cluster heads; sending information to the cluster heads by in-cluster nodes; sending the fused data of the in-cluster nodes between the clusterheads by utilizing a fuzzy rule processing manner to a base station; finally, judging whether conditions of finishing a network life cycle are met or not; after all the nodes in the clusters are dead,carrying out next-time clustering so as to effectively prolong the life cycle of the network, balance the energy consumption of the network and optimize the performance of the network.

The invention relates to a wireless sensor network optimal route path establishing method based on an ant colony algorithm. The wireless sensor network optimal route path establishing method based on the ant colony algorithm is characterized by comprising the steps of (1) randomly electing cluster heads; (2) calculating distances among the cluster heads and distances between all the cluster heads and Sink nodes; (3) establishing a tabu list and an ant prediction energy list, wherein the ant prediction energy list is used for recording current residual energy of all the cluster heads; (4) determining a next hop node, and updating the ant prediction energy list; (5) repeating the step (4) until ants traverse all the nodes, and updating pheromones; (6) executing the step (3) to the step (5) for multiple times in a circulation mode to obtain a dendritic path wherein the dendritic path is the optimal route path. According to the wireless sensor network optimal route path establishing method, the distances of all the sensor nodes in a wireless sensor network and real-time cluster head residual energy are recorded, the global optimal path is found by an ant colony, and thus the performance and the life cycle of the network are improved.

The invention provides a system for compensating an automobile laser radar obstacle recognition blind area, and solves the problem that in the prior art, the blind area exists when a laser radar recognizes an obstacle in a preset range of the automobile. The system comprises a laser radar used for acquiring preset point cloud data information of the obstacle in a preset scanning range, a pluralityof ultrasonic radars used for acquiring preset point data information corresponding to the obstacle in a preset sensing range, and a common space processing module which performs preset time clustering processing on the acquired preset point cloud data and preset point data according to a preset time clustering processing method to acquire obstacle information in the preset sensing range and thepreset scanning range of the current automobile at the same moment. By adopting the system, compensation of blind areas between the laser radar and the ultrasonic radars at different positions at thesame moment can be realized, the blind area of the laser radar obstacle recognition is compensated through few data, and the driving safety of the automobile is further improved.

The invention discloses a method for updating a voiceprint model. The method comprises a step of receiving user voice data, a step of using a zero normalization method to carry out identity authentication on a current user, and taking a user with successful authentication as a target speaker, a step of using a time cluster time to screen voice data of the target speaker and obtaining screened target speaker model updating data, a step of using the target speaker model updating data and original voiceprint model training data to train a new voiceprint model again, and using the new voiceprint model to update the original voiceprint model. According to the method, the continuous false recognition of a speaker similar to the target speaker in a short time is prevented, a condition that the identity authentication of the target speaker can not be carried out caused by the continuous updating of a target speaker voiceprint model is prevented, the inconsistent distribution of model matching degrees caused by different voiceprint model updating degrees of different users is prevented, and thus the accuracy of updating the target speaker voiceprint model is improved.

The invention provides a topology formation and routing method applicable to a wireless optical sensor network. The topology formation and routing method is used for solving technical problems of energy loss imbalance and short network lifetime in the prior art. The method includes steps: electing cluster head nodes according to random numbers generated by nodes; establishing a cluster structure through information interaction of the cluster head nodes, a base station and normal nodes; allowing the cluster head nodes to gather data of the normal nodes and transmit to the base station in single-hop or multi-hop forms; allowing the cluster head nodes to monitor remaining energy at each time slot to establish an optional node set for next-time cluster head election; realizing topology routing characteristics of the wireless optical sensor network. On the basis of an energy model, the cluster head selection optional node set is optimized, and network energy loss is balanced; by introduction of a multi-hop mechanism and a routing mechanism, the problem of limited distance between nodes and the base station is solved, network lifetime is prolonged, and the method is applicable to wireless optical communication sensor networks high in energy balance and lifetime requirements.

A method for automatically classifying images into a final set of events including receiving a first plurality of images having date-time and a second plurality of images with incomplete date-time information; determining one or more time differences of the first plurality of images based on date-time clustering of the images and classify the first plurality of images into a first set of possible events; analyzing the second plurality of images using scene content and metadata cues and selecting images which correspond to different events in the first set of possible events and combining them into their corresponding possible events to thereby produce a second set of possible events; and using image scene content to verify the second set of possible events and to change the classification of images which correspond to different possible events to thereby provide the final set of events.

The invention relates to a spatial clustering method based on time sequence correlation. The method comprises the steps of: 1, selecting a set of spatial points to be clustered; 2, according to geographical relationships of the spatial points, carrying out first-time clustering, and clustering the spatial points belonging to the same geographical relationship into one category; 3, determining a time interval T of time sequence data, which is used in the process of carrying out second-time clustering, obtaining a data value of each spatial point in the time interval T, and forming a time sequence; 4, according to clustering results obtained in the step 2 and the time sequence obtained in the step 3, calculating time sequence correlation between any two spatial points in the same category; and 5, for each clustering result in the step 2, combining the time sequence correlation obtained in the step 4 to carry out second-time clustering on each clustering result so as to form a final clustering result. According to the spatial clustering method disclosed by the invention, two-step clustering is used in the spatial object clustering process, and consideration on the characteristics of time sequence correlation between the objects is added, so that the clustering result is more accurate and has greater practical significance.

The present invention discloses an improved spectral clustering method based on a swarm intelligence algorithm. Feature vectors corresponding to former 2k maximum feature values of a Laplacian matrixare selected as source data of a cluster, and the swarm intelligence algorithm is employed to select high-quality initialization center points for cluster operation so as to improve the maximum accuracy of the cluster and the stability of multi-times cluster results. The Cuckoo search algorithm is introduced to search initialization center points, a fitness function in the process of the Cuckoo search algorithm employs an error square sum function and is applied into the spectral clustering to take data points of the minimum error sum square obtained through search as the initialization centerpoints. The Levy flight strategy in the Cuckoo search algorithm is introduced into the particle swarm optimization algorithm, in a condition that the convergence rate of the particle swarm optimization algorithm is slow, the Levy flight strategy is employed to generate step lengths with small frequencies and occasionally large step lengths, and speed updating formulas for different aspects are introduced in different step lengths.

The invention discloses a maximum information entropy-based chemical data stream real-time abnormity detection method. According to the method, a real-time clustering algorithm DSC-Stream (The Dimension Space Cluster-Stream) is used; a data dimension space is used as a starting point to divide the data dimension space into a plurality of dimension clusters according to the maximum entropy principle; the data in the same dimension cluster forms a micro cluster so as to realize the real-time clustering of the data streams; the algorithm can be used for effectively decreasing the quantity of the invalid grids generated in the traditional density grid algorithm so as to greatly improve the operational efficiency; meanwhile, through a method for setting dimension cluster similarity, the algorithm can be used for effectively solving the problem that the traditional density algorithm is sensitive to the scanning radius and the density threshold to cause poor robustness and can be better adapted to the stream data.

The invention discloses a fine recognition method for urban traffic jam based on mobile clustering. The method comprises the steps: 1, preprocessing track data by carrying out data cleaning and road network matching on the track data and projecting the matched track data to a corresponding time slice; 2, extracting a space cluster with significant high density from each time slice, further measuring the inter-cluster similarity of adjacent time slices, and extracting a candidate jam space-time cluster; and 3) calculating the average speed and the growth duration of the space-time cluster, andif the calculated average speed and the growth duration reach preset conditions, identifying the space-time cluster as a space-time region containing a traffic jam phenomenon. According to the method,the characteristics of fine space-time range, jam scale, survival time and the like of traffic jam in a road network environment can be dug, and the full-life-cycle process of the traffic jam from occurrence to end can be effectively identified based on low-cost vehicle track big data.

The invention discloses a digital PCR concentration calculation method, which determines and calculates the initial concentration of a reactant according to the amplification cycleCt value of the real-time fluorescence quantitative curve in the PCR amplification process combined with the real-time clustering method. The fluorescence threshold R11 has an intersection with the amplification curve ofeach positive reaction unit during exponential growth, and the intersection corresponds to the corresponding amplification cycle value Cti. The digital PCR concentration calculation method specifically includes the following steps: clustering according to the Cti, obtaining k clusters, wherein the central value of each cluster is M1, M2, ..., Mk in descending order; calculating that the number ofthe Cti contained in each cluster is Sj, calculating the number of target genes in each cluster by an Mj and the Sj, adding up the number of target genes in each cluster, finally obtaining the concentration value of the initial target genes.

The invention discloses a big data stream type cluster processing system for on-demand clustering. The system comprises a fast computation module, a data concept drift detection module and a clustering module, wherein an output end of the fast computation module is connected to a first input end of the clustering module through the data concept drift detection module, and the clustering module is connected to the fast computation module. According to the invention, aiming at characteristics of mass, similarity and repetition of the big data, an on-demand clustering model based on data concept drift detection adopts a triggered type clustering processing mode, the accuracy is guaranteed, and on-demand clustering and real-time clustering result services are provided; and secondly, a resource monitoring module and an independent module are provided for clustering processing, the prior traditional clustering algorithms are effectively utilized, expansibility and sensitivity of the system can be enhanced, and quick processing of the data stream in a big data environment is efficiently realized. The big data stream type cluster processing system for on-demand clustering can be widely applied to the field of data processing.

The invention discloses a multi-user clustering and viewpoint calculating system and method based on multiple three-dimensional pictures. The system comprises an infrared capturing module, a position mapping module and a real-time clustering module, wherein the infrared capturing module is used for obtaining the position information of a user, and tracking users in real time, the position mapping module is used for mapping physical coordinates of the infrared capturing stage to the virtual space coordinates, and the real-time clustering module performs real-time dynamic clustering on virtual space users according to user distribution characteristics, takes the central position of a clustering block as viewpoint information of the users, and provides data for follow-up picture rendering. Under the condition that the number of audiences exceeds that of displayed pictures of the system, the audiences with the large position difference can watch the three-dimensional pictures of different viewpoints on the same projection screen, the audiences at the close positions can watch the three-dimensional pictures of the same viewpoint on the same projection screen, and the users can creatively feel stronger and more natural immersed and realistic experience.

The invention provides a man-earth relationship network model and a data processing method. The model comprises a plurality of node units, a plurality of side units and an analyzing unit, wherein each of the node units comprises a kernel module; the kernel module comprises a person kernel module, a place kernel module and an event kernel module which are used for recording relevant information of persons, places, or events respectively; each of the side units comprises a relation side unit and a self-loop side unit; the relation side unit is used for storing and recording relations among persons, places and events among different node units; the self-loop side unit is used for storing and recording change relation of persons, places, or events of the same node unit; and the analyzing unit is used for analyzing comprehensive relations among the persons, the places and the events through a time clustering method or a space clustering method. According to the network model and the data processing method, relations among elements such as persons, places and events can be recorded, so that special problems in special time and space can be analyzed by the time clustering method, a place clustering method, etc.

The invention discloses a fast anomaly detection method for large-scale power transmission and transformation equipment monitoring data streams. The method includes the following steps: performing clustering and category marking on historical data by using a DBScan algorithm, and sending a marked sample to a SparkStreaming cluster to perform real-time incremental clustering; enabling a front-end processor to receive data streams sent by various terminals, pushing the data streams to a Spark Steaming (as shown in the original document) cluster for processing, and implementing real-time featureextraction and normalization processing on the Spark Steaming cluster; and then performing real-time clustering to complete the judgment of a new sample category. By adopting the method, the fast analysis of large-scale data streams of smart power grids can be realized, and the fast anomaly detection of the large-scale power transmission and transformation equipment monitoring data streams can beimplemented.

The invention provides an unmanned plane cluster control system based on a bidirectional communication module and belongs to the technical field of unmanned plane. The bidirectional communication module is integrated in a flight control system, a host plane in a cluster is set, other planes are slave planes, a unique electronic identifier code is set for each unmanned plane, the host plane is operated by an operator, the cluster formation is selected, the target position of each slave plane is calculated through flight control of the host plane according to the real-time self position and therelative position of the slave planes in the formation, identifier codes and the control information are sent to each slave plane through a communication module broadcast mode, after the information is received by the slave planes, the information corresponding to the self identifier codes is selected, read and performed, the self state is fed back to the host plane through the same principle, after data integration by the host plane, the present real-time cluster state is displayed to the operator at the ground control end in a map mode. The system is advantaged in that single-operator unmanned plane cluster control is realized, flight safety and maneuverability are guaranteed through utilizing the host-slave plane mechanism, manpower cost is saved, and great application values are realized in the industrial production field.

The invention provides an aggregation mode discovery method and a system in a road network environment. According to the method, the aggregation mode of motion is discovered by employing motion trackdata of moving objects including vehicles and pedestrians in the road network environment. The method includes: step 1, simplifying tracks, and obtaining similar track segments through the track structure similarity; step 2, comparing characteristics including motion directions and motion speeds of the tracks and recognizing the similar motion tracks through extraction of structural characteristics of the tracks; step 3, clustering the similar track segments through distances, and finding out candidate aggregation mode track sets satisfying a clustering requirement; and step 4, screening a track set in accordance with a space-time condition by employing time clustering, and obtaining an aggregation mode track set. The system comprises three modules including a data preprocessing module, anaggregation mode discovery module and a visualization data analysis display module.