42 results about "Close neighbor" patented technology

The present invention is a system and method for distributed, highly scalable, wide area peer-to-peer network data storage. The functionally equivalent servers in the system are divided into groups. Each server maintains a dynamic list which is polled to determine the availability of the closest neighbor servers. Each server is switched between the groups of servers to optimize network connectivity parameters. Data and directory files are divided into a plurality of pieces which are stored on different servers. Files are uniformly and independently named, utilizing a tree with a common root, logical pathways, and unique file identifiers. When a server receives a client request for file system access, the plurality of file pieces are collected and sent to the client server from the neighbor servers simultaneously in order to optimize bandwidth. Servers with maximum throughput capacity are utilized for highest transmission speed and reduced processing time.

The invention relates to the field of internet communication, and discloses a user group correlation degree-based personalized recommendation method, which comprises the following steps of: A, clustering users by using a clustering algorithm; B, judging distance from a target user to a cluster edge, executing a step C when the distance is greater than a given threshold value, otherwise executing B-1, calculating correlation degree between a cluster of the target user and other clusters, B-2, combining previous r clusters related with the cluster of the user, B-3, searching n closest neighbors in the combined cluster, and further executing a step D; C, searching n closest neighbors in the cluster of the target user; D, predicting a grading value of a related product according to grade of the closest neighbor on the product; and E, selecting the previous m products to be recommended to the user according to the level of a predicted grading value. The invention also discloses a user group correlation degree-based personalized recommendation system. According to the method and the system, the accuracy for personalized recommendation can be effectively improved.

A system and method for creating a peer-to-peer geographical routing overlay network and a multi-attribute similarity routing overlay network. The geographical overlay network can be generated utilizing a geographical routing protocol to organize connections between one or more distributed micro-markets depending on a geographical location of each market and to detect a geographically close neighbor in a decentralized manner. The multi-attribute similarity overlay network can be computed by applying a filtering function based on a multi-attribute routing protocol defined over at least one attribute in order to enable decentralized clustering of the distributed micro-markets. The multi-attribute similarity protocol can be combined with the geographic overlay protocol to route messages to a target set of similarity-based attributes and a target geographic location thereby dynamically evolve a structure of distributed micro-markets over time and optimize a market interaction in the overlay network.

The invention relates to a commodity recommendation method and device based on an electronic business platform and a server. The commodity recommendation method and device based on the electronic business platform and the server are used for improving the accuracy of individual commodity recommendation. The method comprises the steps that adhesive user groups are screened; user interestingness is acquired; a neighbor set closest to a target user is generated; a commodity recommendation list is generated. According to the technical scheme, the obtained adhesive user groups can be more accurate and efficient; due to the fact that only the interestingness of each user in the adhesive user groups is recorded in an obtained scoring matrix, an evaluation matrix can be filled, and the density of the scoring matrix is improved; the closest neighbor set between the target user and the other users in the adhesive user groups is calculated through the scoring matrix, the similar neighbors of the target user can be searched for more accurately, and then the accuracy of recommending commodities to the target user is improved.

An n-ways conferencing system using only the participants' telephony devices and without an external conference server. A plurality of telephony devices dial in to and report their mixing capability to a connection topology manager. The connection topology manager also determines the location of each telephony device and neighboring telephony device pairs in the conference service, as each telephony device is used for mixing voice signals of its closest neighbors, and calculates a minimum conference traffic delay for each telephony device. The connection topology manager then builds a topology for the conference service based on the mixing capability, location, and minimum conference traffic delay, wherein the connection topology manager distributes a mixing computation in a substantially uniform manner across the plurality of telephony devices. When a change in the network condition is detected, the connection topology manager rebuilds the topology to bypass the change in the network condition.

The invention provides a regression prediction method, wherein not only similarity between independent variables X is taken into consideration, but also similarity between dependent variables Y of raw data is taken into consideration, and the model of output value y development based on the historical angle of close neighbors. Compared with the conventional model without taking data development mode into consideration, only one preprocessing section is added to a data set, and the information of the data point can be diversified without the need of extra resource; and the information of the raw data point X is diversified, and finally the prediction effect is improved. Furthermore, the regression prediction method can be realized on a MapReduce frame, and the execution speed can be improved by utilizing the parallelism of the device.

Methods and apparatus teach a digital spectrum of a data file. The digital spectrum is used to map a file's position in multi-dimensional space. This position relative to another file's position reveals closest neighbors. Certain of the closest neighbors are grouped together to define a set. Overlapping members in the groups may be further differentiated from one another by partitioning. An optimized partition of set S of N overlapping groups yields a maximum strength for groups and members in that partition. Among other things, the optimized partition includes relative strengths of every individual member in every possible partition and weighting functions applied to the relative strengths and to subgroups of files within the partitions.

Methods and apparatus teach a digital spectrum of a file representing underlying original data. The digital spectrum is used to map a file's position. This position relative to another file's position reveals closest neighbors. When multiple such neighbors are grouped together they can be used to indicate relevance in current data under consideration on a same or different computing device. Also, relevance can be found without traditional notions of needing structured data or users initiating searching for relevance or by examining metadata/administrative information associated with the files. A plurality of original files represent underlying original bits of data from which a key is created in a mapping space for a relevancy topic. If new data is sufficiently close to this, it is related to the topic and presented to users. Various closeness measures are defined as are methods for key creation. Still other features contemplate computing arrangements and program products.

The invention discloses a separation and recognition algorithm for transformer oiled paper insulation multiple partial discharging source signals. According to the separation and recognition algorithm, time-frequency analysis and close-neighbor similar transmission clusters are adopted to conduct separation and recognition of the multiple partial discharging source signals. The separation and recognition algorithm comprises the steps of firstly, using S transformation to conduct the time-frequency analysis for partial discharging pulses, obtaining an S transformation amplitude (STA) matrix, calculating similarity of the pulses, secondly, using the similarity matrix to conduct the close-neighbor transmission cluster, achieving automatic separation of the discharging pulses, finally, extracting fingerprint feature recognition discharging source modes of a pulse phase partial distribution (PRPD) pattern map of subclass pulses, and using the multiple discharging source signals which are collected through tests and combined by manual work for verification of effectiveness of the separation and recognition algorithm. According to the separation and recognition algorithm for the transformer oiled paper insulation multiple partial discharging source signals, the number of the clusters and corresponding pulse groups can be provided according to the similarity among the pulses and free of influence of the pulse width. When certain pulse width is used for extracting a single pulse wave form for collected PD original data, separation of the multiple discharging sources can be well achieved.

The invention provides a large-scale intelligent group independent collaborative modeling method for imitation of starling group flight. The method comprises the steps of initializing a population andparameters thereof, setting as follows: in the initial state, all Agents use initial positions as optimal positions, and all Agents use optimal values of six or seven most closest neighbors as localoptimal values; calculating a fitness function, and setting distribution anisotropy of neighbors around each Agent, wherein six or seven most closest neighbors are isotropic, and interaction between the Agents depends on a topological distance; selecting an updating company, comprising the step of selecting one of the six or seven most closest neighbors as an updating company for a certain Agent;defining an interaction relationship of Agents; and updating Agent speed and position. The method has a broad application prospect in military, emergency, industrial and medical care fields like dronecluster intensive formation, group emergency evacuation in a large collection place, large-scale robot group collaborative operation, and disease spread control.

An autonomous management cluster of network elements serves as a distributed configuration repository. Network elements sharing a common pre-determined shared identifier autonomously form themselves as a management cluster. The network elements in the cluster exchange configuration files. In the event of a loss, destruction, or corruption of one of the network element's configuration file, the network element recovers its configuration file from its closest neighbor in its management cluster. The management cluster can also be used to efficiently disseminate configuration changes by simply communicating the changes to one or more elements in the cluster, and allowing the other nodes in the cluster to discover and retrieve their updated configuration files.

Methods and apparatus teach a digital spectrum of a file representing underlying original data. The digital spectrum is used to map a file's position. This position relative to another file's position reveals closest neighbors. When multiple such neighbors are grouped together they can be used to indicate relevance in current data under consideration on a same or different computing device. Also, relevance can be found without traditional notions of needing structured data or users initiating searching for relevance or by examining metadata/administrative information associated with the files. A dynamic relevancy agent is configured for installation on the same or different computing device to monitor events regarding the current data and to initiate the examination of relevancy. It also presents to a user suggestions of data closest to the current data. Various triggering events to undertake a relevancy examination are also described as are predetermined criteria to define relative closeness.

A gatekeeper (109) including a compare unit (204) and a record unit (206). The compare unit (204) receives call requests and accesses a database which is a record of where a called party picked up the call, e.g., at home, at the office, via a mobile phone, or via an emergency number. The compare unit (204) performs an analysis such as a “closest neighbor” analysis to determine the location of the user at the time of day closest to that of the present call. The call is then routed to the user at that location. The record unit (206) then makes the appropriate notation in the database for the latest call.

A computer-implemented method, computerized apparatus and computer program product for efficient solution of nearest object problems. A set of objects having non-point shapes representable by one or more geometric elements are obtained. For each object in the set the one or more geometric elements are sampled to obtain a set of points representing the set of objects. A Voronoi diagram is constructed for the set of points, whereby a subdivision of a space accommodating the set of objects into a plurality of cells, each cell containing a single point of the set of points and defining a region of closest neighbors of the point, is obtained. Cells containing points representing the same object are aggregated into a unified cell defining an estimated region of closest neighbors of the object.

The invention relates to an optimized selection method of a user closest neighbor set of an electronic commerce recommendation system. The optimized selection method comprises the steps that user-project scoring data of a historical record is acquired, and a user-project scoring matrix is established; similarity between a target user and other users is calculated by adopting the Pearson similarity calculating method; for the users in a closest neighbor candidate set, an average error between the score of the user and the scores of the users in the set is calculated according to data of a common scoring item set shared by the user and the users in the above mentioned set, and then confidence level is calculated, when the average error is greater, the confidence level between the target user and the users in the set is smaller; K users having the high confidence levels are used to form the closest neighbor set of the target user, and in addition, 1<K<0. Accuracy of recommendation is improved.

Method of detecting a plurality of symbols transmitted by or to a plurality of users from a received signal, each symbol of a user belonging to a modulation constellation, the detection method using a lattice of points generated by the said modulation constellations, the said plurality of symbols of the different users being represented by a point amongst a subset of points in the said lattice, the said constellation and the received signal being represented by a point characteristic of this signal, referred to as the received point, translated from a point in the said constellation by a noise vector, the method including a step of orthogonal projection of the received point onto an affine subspace, referred to as a projection subspace, parallel to or merged with an affine subspace delimiting the constellation, and a step of seeking the closest neighbor to the point thus projected amongst the points in the constellation.

An apparatus includes a capacitive touch screen (CTS); a touch screen interpolator (TSI) coupled to the CTS; a touch screen capacitive memory (TSCM) coupled to the touch screen interpolator, wherein the interpolator is configured to: interpolates a value based on data points correlated to at least three nodes: a magnitude change of capacitance of a node having the largest magnitude change; a position of the largest magnitude of change node; a change of magnitude of capacitance of a first closest neighbor node; and a change of magnitude of capacitance of a second closest neighbor node.

A method of fabricating a photocrystalline plastic optical fiber comprises, for fabricating the cladding, a step of forming a flow by simultaneously injecting a liquid first composition that is a precursor of the cladding polymer and curable by ultraviolet radiation into a first series of holes in an injection plate, and a second composition that is unreactive to the ultraviolet radiation and is selected from a liquid composition and a gas composition into a second series of holes in the plate. The second series of holes have a substantially periodic distribution and each of the holes of the second series has as its closest neighbors holes of the first series. The method further includes a step of irradiating the flow with ultraviolet radiation to form the photocrystalline plastic optical fiber.